Primitives, Listed by Category¶
This page documents all available pipeline primitives, currently 125 in total.
First we list the available primitives in each mode, and below provide the detailed documentation including the parameter arguments for each.
Primitives each have an “order”, which is a floating point number that defines the default ordering when added to a recipe. Smaller numbers come earlier in the execution sequence. You can change the order arbitrarily in the Recipe Editor of course. Notionally, orders <2 are actions on 2D files, orders between 2-4 are actions on datacubes, and orders > 4 are actions on entire observation sequences, but these are not strictly enforced.
(Note: For simplicity, some engineering and software testing related primitives not intended for end users are not listed in the following tables.)
SpectralScience PolarimetricScience Calibration
SpectralScience¶
PolarimetricScience¶
Calibration¶
Primitive Detailed Documentation¶
Display raw image with GPItv¶
Display, with GPItv, raw data to be processed
Category: ALL, HIDDEN Order: 0.01
Inputs: A raw 2D file.
Outputs: No change to data
Notes:
Display in GPITV the current raw image, before any processing
KEYWORDS:
gpitv= session number for the GPITV window to display in.
set to '0' for no display, or >=1 for a display.
HISTORY:
Originally by Jerome Maire 2007-11
2008-04-02 JM: spatial summation window centered on pixel and interpolation on the zem. comm. wav. vector
2008-06-06 JM: adapted to pipeline inputs
2009-04-15 MDP: Documentation updated
2009-09-17 JM: added DRF parameters
2013-07-12 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| gpitv | int | [0,500] | 1 | 1-500: choose gpitv session for displaying output, 0 for no display |
IDL Filename: gpi_display_raw_image_with_gpitv.pro
Correct for Interpixel Capacitance¶
Correct image for interpixel capacitance using Fourier deconvolution.
Category: ALL Order: 0.1
Inputs: Not specified
Outputs: Not specified Output Suffix: ‘ipccor’
Notes:
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
| alpha | float | [0,1] | 0.014 | Fraction of charge in adjacent pixels along columns |
| beta | float | [0,1] | 0.014 | Fraction of charge in adjacent pixels along rows |
IDL Filename: gpi_correct_interpixel_capacitance.pro
Flag Quicklook¶
Flag a given reduction output as ‘quicklook’ quality rather than science grade.
Category: ALL Order: 0.1
Inputs: Not specified
Outputs: The FITS file header in memory gets added a keyword QUIKLOOK=True
Notes:
Writes a QUIKLOOK=True keyword to the current header.
Also updates some FITS history text to indicate the quicklook status.
HISTORY:
Marshall Perrin 2013-10-29 Started based on gpi_add_missingkeyword
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: Save output to disk, 0: Don’t save |
IDL Filename: gpi_flag_quicklook.pro
Load Wavelength Calibration¶
Reads a wavelength calibration file from disk. This primitive is required for any data-cube extraction.
Category: SpectralScience,Calibration Order: 0.5
Inputs: none
Outputs: none; wavecal is loaded into memory
Notes:
Reads a wavelength calibration file from disk.
The wavelength calibration is stored using pointers into the common block.
HISTORY:
Originally by Jerome Maire 2008-07
Documentation updated - Marshall Perrin, 2009-04
2009-09-02 JM: hist added in header
2009-09-17 JM: added DRF parameters
2010-03-15 JM: added automatic detection
2010-08-19 JM: fixed bug which created new pointer everytime this primitive was called
2010-10-19 JM: split HISTORY keyword if necessary
2013-03-28 JM: added manual shifts of the wavecal
2013-04 manual shifts code moved to new update_shifts_for_flexure
2013-07-10 MP: Documentation update and code cleanup
2013-07-16 MP: Rename file for consistency
2013-12-02 JM: get ELEVATIO and INPORT for later flexure correction
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
IDL Filename: gpi_load_wavelength_calibration.pro
Load High-Res PSFs¶
Reads a high-res PSF file from disk. This primitive is required for PSF cube extraction.
Category: PolarimetricScience,SpectralScience,Calibration Order: 0.5
Inputs: none
Outputs: none; mlens psf is loaded into memory
Notes:
Reads a high-res psf file.
The high-res psf is stored using pointers into the common block.
HISTORY:
Originally by Zachary Draper 2-28-14
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
IDL Filename: gpi_load_highres_psfs.pro
Load Polarimetry Spot Calibration¶
Reads a pol spot calibration file from disk. This primitive is required for any polarimetry data-cube extraction.
Category: PolarimetricScience,Calibration Order: 0.51
Inputs: Not used directly
Outputs: none; polarimetry spot cal file is loaded into memory
Notes:
Reads a polarimetry spot calibration file from disk.
The spot calibration is stored using pointers into the common block.
HISTORY:
2013-01-28 MMB: Adapted to pol extraction (based on readwavcal.pro)
2013-02-07 MP: Updated logging and docs a little bit.
Added efficiently not reloading the same file multiple times.
2013-06-04 JBR: shifts for flexure code is now moved to
update_shifts_for_flexure.pro and commented out here.
2013-07-10 MP: Documentation update and code cleanup.
2013-07-17 MP: Rename for consistency
2013-12-16 MP: CalibrationFile argument syntax update.
2014-03-21 MP: Remove 'efficient' code for avoiding reloading, since
this doesn't play well with flexure updates that shift
the calibrations all around.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
IDL Filename: gpi_load_polarimetry_spot_calibration.pro
Load Instrumental Polarization Calibration¶
Load a calibration file for the instrumental polarization.
Category: PolarimetricScience,Calibration Order: 0.52
Inputs: Not specified
Outputs: Instrumental polarization calibration is loaded into memory
Notes:
HISTORY:
2010-05-22 MDP: started
2010-10-19 JM: split HISTORY keyword if necessary
2011-07-30 MP: Updated for multi-extension FITS
2013-07-16 MP: Renamed for consistency
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired instrumental polarization file to be read |
IDL Filename: gpi_load_instrumental_polarization_calibration.pro
Check Data Quality¶
Check quality of data based on header keywords. For bad data, can fail the reduction or simply alert the user.
Category: ALL Order: 0.9
Inputs: 2D image file
Outputs: No change in data; reduction either continues or is terminated.
Notes:
This routine looks at various FITS header keywords to
assess whether the data should be considered usable or not.
The keywords checked include GPIHEALT, AVGRNOT, RMSERR.
You can set the acceptable minimum data quality with the r0 and rmserr
parameters to this primitive.
If a file of unacceptable quality is detected, the action taken will
be determined by the "action" parameter:
0. Alert the user with a message printed to screen
but allow reduction to continue
1. Halt the reduction and fail the receipe.
TODO: implement pop-up window for alerts rather than just
printing a message on screen
GEM/GPI KEYWORDS:AVRGNOT,GPIHEALT,RMSERR
HISTORY:
JM 2010-10 : created
MP 2013-01 : Docs updated
2013-07-16 MP: Documentation cleanup. Rename 'control_data_quality' -> 'check_data_quality'
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Action | int | [0,10] | 1 | 0:Simple alert and continue reduction, 1:Reduction fails |
| r0 | float | [0,2] | 0.08 | critical r0 [m] at lambda=0.5microns |
| rmserr | float | [0,1000] | Critical rms wavefront error in microns. |
IDL Filename: gpi_check_data_quality.pro
Subtract Dark Background¶
Subtract a dark frame.
Category: ALL Order: 1.1
Inputs: raw 2D image file
Outputs: 2D image corrected for dark current Output Suffix: ‘darksub’
Notes:
Subtract background from an image using a dark file.
If CalibrationFile=AUTOMATIC, the best available dark is
obtained from the calibration database.
"Best dark" generally means a dark file that has the most similar
integration time and is closest in date & time of observation
to the data in question.
Specifically, in the Calibration Database code for darks,
the algorithm first looks for dark files which are between
0.3 and 3x of the desired integration time. It takes all such
darks which are on the closest date of observation to the
science data, and from those finds the one that is closest in
integration time to the science data.
This dark is read in, rescaled by the appropriate ratio of
integration times, and then subtracted from the data.
Empirically, rescaling darks by too large a factor does not
result in very high quality subtractions, due to various nonlinear
behaviors such as saturation of hot pixels and the so-called
'reset anomaly' effect which biases the readout background level.
Hence we impose a limit for scaling the dark integration time
up or down, semi-arbitrarily chosen to be 3x because it seems to
work reasonably well. The standard set of darks planned to be
taken routinely at Gemini should ensure that there are always available
darks within this range.
If you desire different behavior, simply set the CalibrationFile manually
of course.
Note: If the RequireExactMatch setting is 1, then only dark files
exactly matching in integration time will be used. If there is no
such file, the data is returned without any subtraction.
HISTORY:
Originally by Jerome Maire 2008-06
2009-04-20 MDP: Updated to pipeline format, added docs.
Some code lifted from OSIRIS subtradark_000.pro
2009-09-02 JM: hist added in header
2009-09-17 JM: added DRF parameters
2010-10-19 JM: split HISTORY keyword if necessary
2012-07-20 MP: added DRPDARK keyword
2012-12-13 MP: Remove "Sky" from primitve discription since it's inaccurate
2013-07-11 MP: rename 'applydarkcorrection' -> 'subtract_dark_background' for consistency
2013-10-03 MP: Add RequireExactMatch option, enable scaling for non-matching exptimes
2013-12-16 MP: CalibrationFile argument syntax update.
2014-03-22 MP: Adding experimental interpolation option.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Name of dark file to subtract |
| RequireExactMatch | int | [0,1] | 0 | Must dark calibration file exactly match in integration time, or is scaling from a different exposure time allowed? |
| Interpolate | int | [0,1] | 0 | Interpolate based on JD between prior and subsequent available darks |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_subtract_dark_background.pro
Subtract Thermal/Sky Background if K band¶
Subtract a dark frame.
Category: ALL Order: 1.2
Inputs: 2D image file
Outputs: 2D image file, unchanged if YJH, background subtracted if K1 or K2. Output Suffix: ‘bkgndsub’
Notes:
Subtract thermal background emission, for K band data only
** special note: **
This is a new kind of "data dependent optional primitive". If the filter of
the current data is YJH, return without doing *anything*, even logging the
start/end of this primitive. It becomes a complete no-op for non-K-band
cases.
Algorithm:
Get the best available thermal background calibration file from CalDB
Scale it to current exposure time
Subtract it.
The name of the calibration file used is saved to the DRPBKGND header keyword.
ALGORITHM TODO: Deal with uncertainty and pixel mask frames too.
HISTORY:
2012-12-13 MP: Initial implementation
2013-01-16 MP: Documentation cleanup.
2013-07-12 MP: Rename for consistency
2013-12-15 MP: Add override_scaling option, remove erroneous hard-coded
constant non-1 scaling.
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Name of thermal background file to subtract |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| Override_scaling | float | [0,10] | 1.0 | Set to value other than 1 to manually adjust the background image flux scaling to better match the science data |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_subtract_thermal_sky_background_if_k_band.pro
Remove Persistence from Previous Images¶
Determines/Removes persistence of previous images
Category: ALL Order: 1.2
Inputs: Raw or destriped 2D image
Outputs: 2D image corrected for persistence of previous non-saturated images Output Suffix: ‘-nopersis’
Notes:
The removal of persistence from previous non-saturated images
incorporates a model developed for Hubble Space Telescopes Wide
Field Camera 3 (WFC3,
www.stsci.edu/hst/wfc3/ins_performance/persistence).
Persistence is proportional to the intensity of the illuminating
source, and is observed to fade exponentially with time. The
parameters of the mathematical model for the persistence, found
in the pipeline's configuration directory were determined
during integration and test at UCSC.
This primitive searches for all files in the raw data directory
taken within 600 seconds (10 min) of the beginning of the exposure
of interest. It then calculates the persistence from each image,
using the maximum of the stack, and subtracts it from the
frame. Note that if the detector is exposed to light, but no
exposures are being taken, persistence will still build up on the
detector that cannot be subtracted.
Ideally, this program should be run after the destriping algorithm
as readnoise does not induce persistence. However, due to limitation
that a pipeline primitive cannot call another primitive, this has
not been implemented. Future developement will involve moving the
destriping algorithm into a idl function, and then calling the
function from the destriping primitive. This will enable the ability
for this primitive to destripe the previous images. The user should
note that the destriping is at a level that is low enough to not
leave a significant persistence, so this detail will not
significantly affect science data.
At this time, the persistence is removed at the ~75% level due to
inaccuracies in the model caused by an insufficient time sampling of
the initial falloff and readnoise. A new dataset will be taken prior to shipping,
and new model parameters will be derived prior to commissioning.
WARNING: Persistence removal does not (yet) work with COADDED images!
The manual_UTEND keyword allows the user to manually set the UTEND keyword in the image that is taken in closest time to the image being reduced. This is important because sometimes when changing modes or exposure times quickly, the CAL exit shutter can remain open so light will continue hitting the detector past the UTEND time. Users should note that this is generally occurs only for polarimetry snapshots after long spectral sequences (taken in the same band).
Requires the persistence_model_parameters.fits calibration file.
HISTORY:
Wed May 22 15:11:10 2013, LAB <LAB@localhost.localdomain>
2013-05-14 PI: Started
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the persistence_parameter file to be read |
| manual_dt | float | [0,600] | 0 | Manual input for time (in seconds) since last persisting file - see help for details |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_remove_persistence_from_previous_images.pro
Clean Cosmic Rays¶
Placeholder for cosmic ray rejection (if needed; not currently implemented!)
Category: HIDDEN Order: 1.23
Inputs: Not specified
Outputs: Not specified
Notes:
Placeholder; does not actually do anything yet.
Empirically, cosmic rays do not appear to be a significant noise source
for the GPI IFS. It's a substrate-removed H2RG so the level is quite low.
Furthermore, realtime identification and removal of CRs is included as
part of the up-the-ramp readout and slope fitting, which handles the
majority of CRs.
There are still occasional noticeable residual CRs, particularly in long
duration exposures or darks, but they've not yet proven annoying enough to
implement an algorithm here...
HISTORY:
2010-01-28 MDP: Created Templae.
2011-07-30 MDP: Updated for multi-extension FITS
2013-07-16 MDP: Renamed as part of code cleanup.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_clean_cosmic_rays.pro
Apply Reference Pixel Correction¶
Subtract channel bias levels and bias drift stripes using H2RG reference pixels.
Category: ALL Order: 1.25
Inputs: 2D image file
Outputs: 2D image corrected for background using reference pixels Output Suffix: ‘refpixcorr’
Notes:
Correct for fluctuations in the bias/dark level using the rows of
reference pixels in the H2RG detectors.
Note that *vertical* reference pixel subtraction to fix offsets between
the 32 readout channels is done in real time during the readout process by
the IFS Detector Server software. The Detector Server does not currently
apply any horizontal reference pixel subtraction, so we need to do that in
the pipeline. See the HRPSTYPE and VRPSTYPE FITS keywords in the SCI
extension headers.
Also note that if you use one of the specialized Destriping primitives,
you do not also need to use this one as well.
Algorithm choices include:
1) simple_channels in this case, just use the median of each
vertical channel to remove offsets between
the channels. (deprecated, now done by the IFS
detector server in real time during readout)
2) simple_horizontal take the median of the 8 ref pix for each row,
and subtract that from each row.
3) smoothed_horizontal Like the above, but smoothed by N pixels vertically
for better S/N. N is adjustable using the smoothing_size
parameter. Empirically values < 20 or 30 seem to be
not enough smoothing, so the read noise fluctuations
give spurious biases to the ref pix model.
3) interpolated In this case, use James Larkin's interpolation
algorithm to remove linear variation with time
in the horizontal direction. This gives the highest
spatial frequency correction but is more affected
by read noise.
See discussion in section 3.1 of Rauscher et al. 2008 Prof SPIE 7021 p 63.
ALGORITHM TODO: Deal with uncertainty and pixel mask frames too.
HISTORY:
Originally by Jerome Maire 2008-06
2009-04-20 MDP: Updated to pipeline format, added docs.
Some code lifted from OSIRIS subtradark_000.pro
2009-09-17 JM: added DRF parameters
2012-07-27 MP: Added Method parameter, James Larkin's improved algorithm
2012-10-14 MP: debugging and code cleanup.
2013-07-17 MP: Rename for consistency
2013-12-03 MP: Some docs updates and added SMOOTHED_HORIZONTAL algorithm and smoothing_size parameter
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
| smoothing_size | int | [0,500] | 31 | Smoothing kernel size for smoothed_horizontal method. |
| before_and_after | int | [0,1] | 0 | Show the before-and-after images for the user to see? |
| Method | enum | SIMPLE_CHANNELS|SIMPLE_HORIZONTAL|SMOOTHED_HORIZONTAL|INTERPOLATED | INTERPOLATED | Algorithm for reference pixel subtraction. |
IDL Filename: gpi_apply_reference_pixel_correction.pro
Update Spot Shifts for Flexure¶
Extract a 3D datacube from a 2D image. Spatial integration (3 pixels) along the dispersion axis
Category: SpectralScience, Calibration, PolarimetricScience Order: 1.34
Inputs: Not specified
Outputs: Not specified Output Suffix: Could not be determined automatically
Notes:
This primitive updates the wavelength calibration and spot location table
to account for shifts in the apparent position of each spectrum due to
elevation-dependent flexure within the IFS. The observed image motion is
about 0.7 pixels in X and 0.5 pixels in Y between 0 and 90 degrees
By updating the X and Y coordinates of each lenslet across the field of view,
this primitive enables the extraction of well behaved data cubes
regardless of the orientation.
There are several options for how to determine the shifts, set by the
method keyword:
method="None" No correction applied.
method='Manual' Apply shifts provided by the user via the
manual_dx and manual_dy arguments.
method='Lookup' Correction applied based on a lookup table of shifts
precomputed based on arc lamp data at multiple
orientations, obtained from the calibration
database.
method='BandShift'Estimate the flexure values by comparing to the
most recent wavecal regardless of the band and
interpolating.
method='Auto' [work in progress, use at your own risk]
Attempt to determine the shifts on-the-fly from each
individual exposure via model fitting.
If the 'gpitv' argument to this primitive is used to send the output
image to a gpitv session, it will be displayed *with the updated
wavelength calibration information overplotted*.
HISTORY:
2013-03-08 MP: Started based on extractcube, initial attempts at automated
on-the-fly measurements.
2013-03-25 JM: Implemented lookup table version.
2013-04-22 PI: A few bug fixes to lookup table code.
2013-04-25 MP: Documentation improvements.
2013-06-04 JBR: Now compatible with polarimetry.
2013-07-17 MP: Rename for consistency
2013-12-02 JM: new way of dealing with the lookup table for flexure effect correction, independent of the reference wavelength solution used to calculate the shifts
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| method | string | [None|Manual|Lookup|BandShift|Auto] | None | How to correct spot shifts due to flexure? [None|Manual|Lookup|BandShift|Auto] |
| manual_dx | float | [-10,10] | 0 | If method=Manual, the X shift of spectra at the center of the detector |
| manual_dy | float | [-10,10] | 0 | If method=Manual, the Y shift of spectra at the center of the detector |
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window; else = Window number to display diagnostic plot. |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_update_spot_shifts_for_flexure.pro
Flexure 2D x correlation with polcal¶
This primitive uses the relevent pol cal file to generate a model detector image to cross correlate with a science image and find the flexure offset.
Category: PolarimetricScience Order: 1.34
Inputs: Science image, polcal
Outputs: Flexure offset in xy detector coordinates. Output Suffix: ‘’ ; set this to the desired output filename suffix
Notes:
This primitive uses the relevent microlense PSF and pol cal to generate a model detector image to cross correlate with a science image.
The resulting output can be used as a flexure offset prior to flux extraction.
The resulting output can be used as a flexure offset prior to flux extraction.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
2014-09-12 MMB: Branched to a version that cross correlates with the polcal file rather than use the microlens
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| range | float | [0,5] | 0.3 | Range of cross corrleation search in pixels. |
| resolution | float | [0,1] | 0.1 | Subpixel resolution of cross correlation |
| psf_sep | float | [0,1] | 0.1 | PSF separation in pixels |
| stopidl | int | [0,1] | 0 | 1: stop IDL, 0: dont stop IDL |
| x_off | float | [-5,5] | 0 | initial guess for large offsets |
| y_off | float | [-5,5] | 0 | initial guess for large offsets |
| badpix | float | [0,1] | 1 | Weight by bad pixel map? |
| iterate | int | [0,1] | 1 | Take the first result? Or iterate |
| max_iter | int | [1,100] | 15 | The maximum number of iterations |
IDL Filename: gpi_img_xcorr_polcal.pro
Destripe for Darks Only¶
Subtract readout pickup noise using median across all channels. This is an aggressive destriping algorithm suitable only for use on images that have no light. Also includes microphonics noise removal.
Category: Calibration Order: 1.35
Inputs: A 2D dark image
Outputs: 2D image corrected for stripe noise Output Suffix: ‘destripe’
Notes:
Correct for fluctuations in the background bias level
(i.e. horizontal stripes in the raw data) using a pixel-by-pixel
median across all channels, taking into account the alternating readout
directions for every other channel.
This provides a very high level of rejection for stripe noise, but of course
it assumes that there's no signal anywhere in your image. So it's only
good for darks.
A second noise source that can be removed by this routine is the
so-called microphonics noise induced by high frequency vibrational modes of
the H2RG. This noise has a characteristic frequenct both temporally and
spatially, which lends itself to removal via Fourier filtering. After
destriping, the image is Fourier transformed, masked to select only the
Fourier frequencies of interest, and transformed back to yield a model for
the microphonics striping that can be subtracted from the data. Empirically
this correction works quite well. Set the "remove_microphonics" option to
enable this, and set "display" to show on screen a
diagnostic plot that lets you see the stripe & microphonics removal in
action.
SEE ALSO: Destripe science frame
HISTORY:
2012-10-16 Patrick: fixed syntax error (function name)
2012-10-13 MP: Started
2013-01-16 MP: Documentation cleanup
2012-03-13 MP: Added Fourier filtering to remove microphonics noise
2013-04-25 MP: Improved documentation, display for microphonics removal.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| remove_microphonics | string | [yes|no] | yes | Attempt to remove microphonics noise via Fourier filtering? |
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window else = Window number to display diagonostics in. |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_destripe_for_darks_only.pro
Destripe science image¶
Subtract detector striping using measurements between the microspectra
Category: SpectralScience,Calibration, PolarimetricScience Order: 1.35
Inputs: Not specified
Outputs: Not specified Output Suffix: Could not be determined automatically
Notes:
This primitive was originally developed to remove striping and microphonics
noise in IFS images. The noise level of the detector has since decreased
significantly and therefore this primitive is generally only useful for
short exposures. Note that without proper examination, this primitive may
INTRODUCE a systematic noise into the image. Users should consult the
IFS handbook destriping section when using this primitive.
This primitive subtracts horizontal striping from the background of a 2d raw IFS image
by masking spectra and using the remaining regions to obtain a
sampling of the striping.
The masking can be performed by using the wavelength calibration to mask the
spectra (recommended) or by thresholding (not recommended).
WARNING: This destriping algorithm will not work correctly on flat fields or
any image where there is very large amounts of signal covering the entire
field. If called on such data, it will print a warning message and return
without modifying the data array.
Summary of the primitive:
The principle idea is to build models of the different source of noise
you want to treat and then subtract them from the real image at the end.
1/ mask computation
2/ Channels offset model based on im = image => chan_offset
3/ Microphonics computation based on im = image - chan_offset => microphonics_model
4/ Destriping model based on im = image - chan_offset - microphonics_model => stripes
5/ Output: imout = image - chan_offset - microphonics_model - stripes
Destriping Algorithm Details:
Generate a mask of where the spectra are located, based on the
already-loaded wavelength or pol spots solutions.
Mask out those pixels.
Break the image up into the 32 readout channels
Flip the odd channels to account for the alternating readout direction.
Generate a median image across the 32 readout channels
Smooth by 20 pixels to generate the broad variations
mask out any pixels that are >3 sigma discrepant vs the broad variations
Generate a better median image across the 32 readout channels post masking
Perform some sanity checks for model validity and interpolate NaNs as needed
Expand to a 2D image model of the detector
OPTIONAL/EXPERIMENTAL:
The microphonics noise attenuation can be activitated by setting the parameter remove_microphonics to 1 or 2.
The microphonics from the image can be saved in a file using the parameter save_microphonics.
If Plot_micro_peaks equal 'yes', then it will open 3 plot windows with the peaks aera of the
microphonics in Fourier space (Before microphonics subtraction, the
microphonics to be removed and the final result). Used for debugging purposes.
If remove_microphonics = 1:
The algorithm is always applied.
If remove_microphonics = 2:
The algorithm is applied only of the quantity of noise is greater than the micro_threshold parameter.
A default empirical value of 0.01 has been set based on the experience of the author of the algorithm.
The quantity of microphonics noise is measured with the ratio of the dot_product and the norm of the
image: dot_product/sqrt(sum(abs(fft(image))^2)).
With dot_product = sum(abs(fft(image))*abs(fft(noise_model))) which
correspond to the projection of the image on the microphonics noise model in the absolute Fourier space.
There are 3 implemented methods right now depending on the value of the parameter method_microphonics.
If method_microphonics = 1:
The microphonics noise removal is based on a fixed precomputed model. This model is the
normalized absolute value of the Fourier coefficients.
The filtering consist of diminishing the intensity of the frequencies corresponding to the
noise in the image proportionaly to the dot product of the image witht the noise model.
The phase remains unchanged.
The filtered coefficients in Fourier space become (1-dot_product*(Amplitude_noise_model/Amplitude_image)).
With dot_product = sum(abs(fft(image))*abs(fft(noise_model))) which correspond to the projection of the image on the microphonics noise model in the absolute Fourier space.
If method_microphonics = 2:
The frequencies around the 3 identified peaks of the microphonics noise in Fourier space are all set to zero.
This algorithm is the best one of you are sure that there is no data in this aera but it is probably better not to use it...
If method_microphonics = 3:
A 2d gaussian is fitted for each of the three peaks of the microphonics noise in Fourier space and then removed.
Only the absolute value is considered and the phase remains unchanged.
This algorthim is not as efficient as the two others but if you don't have an accurate model, it can be better than nothing.
Currently, the readnoise floor, which is what is used to determine the pixel masking for spectral mode, is set to 8 electrons divided by the
square root of hte number of coadds. Note that for K band (and sometimes H) this often has to be adjusted. The channel
offset correction should also be used when this value is being adjusted. Note that if too much of the image is masked,
it will surpass the abort_fraction and no destriping will occur. Using an abort_fraction of 0.7 is the minimum
a user should use for normal cases.
HISTORY:
Originally by Marshall Perrin, 2011-07-15
2011-07-30 MP: Updated for multi-extension FITS
2012-12-12 PI: Moved from Subtract_2d_background.pro
2012-12-30 MMB: Updated for pol extraction. Included Cal file, inserted IDL version checking for smooth() function
2013-01-16 MP: Documentation cleanup.
2013-03-12 MP: Code cleanup, some speed enhancements by vectorization
2013-05-28 JBR: Primitive copy pasted from the destripe_mask_spectra.pro primitive. Microphonics noise enhancement. Microphonics algorithm now applied before the destriping.
2013-12-04 PI: Removed high_limit- now does masking based on readnoise levels
2013-12-30 MP: CalibrationFile argument syntax update.
2014-02-25 MP: flats in polarization mode are OK to destripe
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| method | string | [threshold|calfile] | calfile | Find background based on image value threshold cut, or calibration file spectra/spot locations? |
| abort_fraction | float | [0.0,1.0] | 0.9 | Necessary fraction of pixels in mask to continue - set at 0.9 to ensure quicklook tool is robust |
| chan_offset_correction | int | [0,1] | 0 | Tries to correct for channel bias offsets - useful when no dark is available |
| readnoise_floor | float | [0.0,100] | 0.0 | Readnoise floor in ADU. 0 = default to 8 electrons per CDS image |
| Save_stripes | int | [0,1] | 0 | Save the striping noise image subtracted from frame? |
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window else = Window number to display diagonostics in. |
| remove_microphonics | int | [0,2] | 0 | Remove microphonics noise based on a precomputed fixed model.0: not applied. 1: applied. 2: the algoritm is applied only if the measured noise is greater than micro_threshold |
| method_microphonics | int | [1,3] | 0 | Method applied for microphonics 1: model projection. 2: all to zero 3: gaussian fit |
| CalibrationFile | string | None | AUTOMATIC | Filename of the desired microphonics model file to be read |
| Plot_micro_peaks | string | [yes|no] | no | Plot in 3d the peaks corresponding to the microphonics |
| save_microphonics | string | [yes|no] | no | If remove_microphonics = 1 or (auto and micro_threshold overpassed), save the removed microphonics |
| micro_threshold | float | [0.0,1.0] | 0.01 | If remove_microphonics = 2, set the threshold. This value is sum(abs(fft(image))*abs(fft(noise_model)))/sqrt(sum(image^2)) |
| write_mask | int | [0,1] | 0 | write signal mask to reduced directory? |
| fraction | float | [0.0,1.0] | 0.7 | Threshold fraction of the total pixels in a row should be masked |
| Save | int | [0,1] | 0 | 1: Save output to disk, 0: Don’t save |
| gpitv | int | [0,500] | 1 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_destripe_science_image.pro
Interpolate bad pixels in 2D frame¶
Repair bad pixels by interpolating between their neighbors. Can optionally just flag as NaNs or else interpolate.
Category: SpectralScience, PolarimetricScience, Calibration Order: 1.4
Inputs: 2D image, ideally post dark subtraction and destriping
Outputs: 2D image with bad pixels marked or cleaned up. Output Suffix: ‘-bpfix’
Notes:
Interpolates between vertical (spectral dispersion) direction neighboring
pixels to fix each bad pixel.
Bad pixels are identified from:
1. The pixels marked bad in the current bad pixel mask (provided in the
CalibrationFile parameter.)
2. Any additional pixels which are marked as bad in the image extension
for data quality (DQ).
3. Any pixels which are < -50 counts (i.e. are > 5 sigma negative where
sigma is the CDS read noise for a single read). TODO: This threshhold
should be evaluated and possibly made adjustible.
The action taken on those bad pixels is determined from the 'method'
parameter, which can be one of:
'nan': Bad pixels are just marked as NaN, with no interpolation
'vertical': Bad pixels are repaired by interpolating over their
immediate neighbors vertically, the pixels above and below.
This has been shown to work well for spectral mode GPI data
since vertical is the spectral dispersion direction.
(The actual algorithm is a bit more complicated than this to
handle cases where the above and/or below pixels are themselves
also bad.)
'all8': Repair by interpolating over all 8 surrounding pixels.
TODO: need to evaluate whether that algorithm is still a good approach for
polarimetry mode files.
TODO: implement Christian's suggestion of a 3D interpolation in 2D space,
using adjacent lenslet spectra as well. See emails of Oct 18, 2012
(excerpted below)
HISTORY:
Originally by Marshall Perrin, 2012-10-18
2012-12-03 MP: debugging/enhancements for the case of multiple adjacent bad
pixels
2012-12-09 MP: Added support for using information in DQ extension
2013-01-16 MP: Documentation cleanup
2013-02-07 MP: Enhanced all8 interpolation to properly handle cases where
there are bad pixels in the neighboring pixels.
2013-04-02 JBR: Correction of a sign in the vertical algorithm when reading the bottom adjacent pixel.
2013-04-22 JBR: In vertical algorithm, condition added if both upper and bottom pixels are good.
2013-06-26 MP: Added better FITS history logging for the case of not having a bad pixel map.
2013-07-12 MP: Rename file for consistency
2013-12-16 MP: Update to allow bad pixel map files to have values other than
1, with any nonzero value being interpreted as bad.
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | None | None | None | Filename of the desired bad pixel file to be read |
| method | string | [n4n|vertical|all8] | vertical | Repair bad bix interpolating all 8 neighboring pixels, or just the 2 vertical ones, or just flag as NaN (n4n)? |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 1 | 1-500: choose gpitv session for displaying output, 0: no display |
| negative_bad_thresh | float | [-100000,0] | -50 | Pixels more negative than this should be considered bad. |
| before_and_after | int | [0,1] | 0 | Show the before-and-after images for the user to see? (for debugging/testing) |
IDL Filename: gpi_interpolate_bad_pixels_in_2d_frame.pro
Combine 2D images¶
Combine 2D images such as darks into a master file via mean or median.
Category: ALL Order: 1.5
Inputs: Multiple 2D images
Outputs: a single combined 2D image Output Suffix: strlowcase(method)
Notes:
Multiple 2D images can be combined into one using either a mean,
a sigma-clipped mean, or a median.
HISTORY:
Jerome Maire 2008-10
2009-09-17 JM: added DRF parameters
2009-10-22 MDP: Created from mediancombine_darks, converted to use
accumulator.
2010-01-25 MDP: Added support for multiple methods, MEAN method.
2011-07-30 MP: Updated for multi-extension FITS
2012-10-10 MP: Minor code cleanup
2013-07-10 MP: Minor documentation cleanup
2013-07-12 MP: file rename for consistency
2014-01-02 MP: Copied SIGMACLIP implementation from gpi_combine_2d_dark_images
2014-11-04 MP: Avoid trying to run parallelized sigmaclip if in IDL runtime.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | string | MEAN|MEDIAN|SIGMACLIP | SIGMACLIP | How to combine images: median, mean, or mean with outlier rejection?[MEAN|MEDIAN|SIGMACLIP] |
| Sigma_cut | float | [1,100] | 3 | If Method=SIGMACLIP, then data points more than this many standard deviations away from the median value of a given pixel will be discarded. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_2d_images.pro
Combine 2D Thermal/Sky Backgrounds¶
Combine 2D images with measurement of thermal or sky background
Category: Calibration Order: 1.51
Inputs: 2D image(s) taken with lamps off.
Outputs: thermal background file, saved as calibration file Output Suffix: Could not be determined automatically
Notes:
Generate a 2D background image for use in removing e.g. thermal emission
from lamp images
HISTORY:
2012-12-13 MP: Forked from combine2dframes
2013-07-10 MP: Minor documentation cleanup
2013-07-12 MP: Rename for consistency
2014-01-02 MP: Copied SIGMACLIP implementation from gpi_combine_2d_dark_images
2014-11-10 MP: Avoid trying to run parallelized sigmaclip if in IDL runtime.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | enum | MEAN|MEDIAN|SIGMACLIP | SIGMACLIP | How to combine images: median, mean, or mean with outlier rejection?[MEAN|MEDIAN|SIGMACLIP] |
| Sigma_cut | float | [1,100] | 3 | If Method=SIGMACLIP, then data points more than this many standard deviations away from the median value of a given pixel will be discarded. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_2d_thermal_sky_backgrounds.pro
2D Wavelength Solution Developer¶
This primitive uses an existing wavelength solution file to construct a new wavelength solution file by simulating the detector image and performing a least squares fit.
Category: Calibration Order: 1.7
Inputs: An Xe/Ar lamp detector image
Outputs: A wavelength solution cube (and a simulated Xe/Ar lamp detector image; to come) Output Suffix: ‘wavecal’
Notes:
This is the main wavelength calibration generation primitive.
This Wavelength Solution generator models an arclamp spectrum
for each lenslet and uses mpfit2dfunc to fit the relevant
wavelength solution variables (ie. xo, yo, lambdao, dispersion,
tilt). A wavelength solution file is output along with a
simulated detector image.
A previous wavelength calibration file is used to supply the
initial guess for the fitting process, which is then updated
by this primitive.
This is fairly computationally intensive and requires
relatively high S/N data. See Quick Wavelength Solution if
you need faster results (albeit more limited and requiring you
already have a reference wavecal)
HISTORY:
2013-09-19 SW: 2-dimensionsal wavelength solution
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| display | Int | [0,1] | 0 | Whether or not to plot each lenslet spectrum model in comparison to the detector measured spectrum: 1;display, 0;no display |
| whichpsf | Int | [0,1] | 0 | Type of lenslet PSF model, 0: gaussian, 1: microlens |
| parallel | Int | [0,1] | 0 | Option for Parallelization, 0: none, 1: parallel |
| numsplit | Int | [0,100] | 0 | Number of cores for parallelization. Set to 0 for autoselect. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| Save_model_image | int | [0,1] | 0 | 1: save 2d detector model fit image to disk, 0:don’t save |
| CalibrationFile | wavcal | None | AUTOMATIC | Filename of the desired reference wavelength calibration file to be read |
| Save_model_params | int | [0,1] | 0 | 1: save model nuisance parameters to disk, 0: don’t save |
| AutoOffset | int | [0,1] | 0 | Automatically determine x/yoffset values 0;NO, 1;YES |
IDL Filename: gpi_wavelength_solution_2d_developer.pro
Quick Wavelength Solution Update¶
Given an existing wavecal and a new Xe lamp image, this primitive updates the wavecal based on the X,Y positions measured for a subset of the Xe spectra.
Category: Calibration Order: 1.7
Inputs: An Xe/Ar lamp detector image
Outputs: Not specified Output Suffix: ‘wavecal’
Notes:
This is a modified version of the 2D wavelength solution
algorithm, which fits a small subset of lenslets (set by
the 'spacing' argument) to very quickly provide an estimated
wavelength solution, based on some prior wavelength solution.
This differs from the full wavelength solution in that:
1) Only a subset of lenslets are fit
2) The mean shifts in X and Y are derived from those fits
3) The output wavelength solution is created by taking
the input wavelength solution and applying those shifts.
(i.e. only the overall shift of the wavecal is updated;
the individual dispersions and tilts of each lenslet's
spectrum are not changed).
This algorithm is both computationally faster than and
tolerant of lower S/N data than the full wavelength solution
algorithm. This is because it is in essence only trying to measure
2 parameters, the average shifts in X and Y, rather than the
~ 150,000 parameters measured and saved for the full wavelength
calibration algorithm.
KEYWORDS:
GEM/GPI KEYWORDS:FILTER,IFSFILT,GCALLAMP
DRP KEYWORDS: FILETYPE,HISTORY,ISCALIB
HISTORY:
2013-09-19 SW: 2-dimensionsal wavelength solution
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window; else = Window number to display each lenslet in comparison to the detector lenslet in. |
| spacing | Int | [0,20] | 10 | Test every Nth lenslet for this value of N. |
| boxsizex | Int | [0,15] | 7 | x dimension of a lenslet cutout |
| boxsizey | Int | [0,50] | 24 | y dimension of a lenslet cutout |
| xoffset | Int | [-10,10] | 0 | x offset guess from prior wavecal. |
| yoffset | Int | [-20,20] | 0 | y offset guess from prior wavecal. |
| whichpsf | Int | [0,1] | 0 | Type of psf 0;gaussian, 1;microlens |
| high_order_correction | Int | [0,1] | 1 | Higher order flexure offsets? 0; No, 1; Yes |
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| AutoOffset | int | [0,1] | 0 | Automatically determine x/yoffset values 0;NO, 1;YES |
| gpitvim_dispgrid | int | [0,500] | 15 | 1-500: choose gpitv session for displaying image output and wavcal grid overplotted, 0: no display |
IDL Filename: gpi_quick_wavelength_solution_update.pro
2D Wavelength Solution¶
This primitive uses an existing wavelength solution file to construct a new wavelength solution file by simulating the detector image and performing a least squares fit.
Category: Calibration Order: 1.7
Inputs: An Xe/Ar lamp detector image
Outputs: A wavelength solution cube (and a simulated Xe/Ar lamp detector image; to come) Output Suffix: ‘wavecal’
Notes:
This is the main wavelength calibration generation primitive.
This Wavelength Solution generator models an arclamp spectrum
for each lenslet and uses mpfit2dfunc to fit the relevant
wavelength solution variables (ie. xo, yo, lambdao, dispersion,
tilt). A wavelength solution file is output along with a
simulated detector image.
A previous wavelength calibration file is used to supply the
initial guess for the fitting process, which is then updated
by this primitive.
This is fairly computationally intensive and requires
relatively high S/N data. See Quick Wavelength Solution if
you need faster results (albeit more limited and requiring you
already have a reference wavecal)
HISTORY:
2013-09-19 SW: 2-dimensionsal wavelength solution
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| display | Int | [0,1] | 0 | Whether or not to plot each lenslet spectrum model in comparison to the detector measured spectrum: 1;display, 0;no display |
| parallel | Int | [0,1] | 0 | Option for Parallelization, 0: none, 1: parallel |
| numsplit | Int | [0,100] | 0 | Number of cores for parallelization. Set to 0 for autoselect. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| Save_model_image | int | [0,1] | 0 | 1: save 2d detector model fit image to disk, 0:don’t save |
| CalibrationFile | wavcal | None | AUTOMATIC | Filename of the desired reference wavelength calibration file to be read |
| Save_model_params | int | [0,1] | 0 | 1: save model nuisance parameters to disk, 0: don’t save |
IDL Filename: gpi_wavelength_solution_2d.pro
Measure Wavelength Calibration¶
Derive wavelength calibration from an arc lamp or flat-field image.
Category: Calibration Order: 1.7
Inputs: 2D image from narrow band arclamp
Outputs: Output Suffix: Could not be determined automatically
Notes:
This primitive positions of spectra in the image with narrow
band lamp image.
** DEPRECATED** This is the older 'first generation' wavelength
calibration algorith, which is no longer recommended
ALGORITHM:
gpi_extract_wavcal starts by detecting the central peak of the image.
Next, starting with a initial value of w & P, find the nearest peak (with an increment on the microlens coordinates)
when nearest peak has been detected, it reevaluates w & P and so forth..
*********************************************************************************
*
* IMPORTANT WARNING for future software maintainers:
* The complicated algorithms implemented here were originally developed
* assuming the dispersion direction in GPI would be horizontal. Given data
* orientation conventions later adopted, it became vertical. Rather than
* rewriting all of the following and swapping all the indices around,
* the images are just *transposed* as the first step of this process, and
* then the original horizontal algorithm applied. This leads to various
* complexities about index transformations. Be wary when editing the
* code here and keep that in mind....
*
*
*********************************************************************************
common needed:
KEYWORDS:
GEM/GPI KEYWORDS:FILTER,IFSFILT,GCALLAMP,GCALSHUT,OBSTYPE
DRP KEYWORDS: FILETYPE,HISTORY,ISCALIB
HISTORY:
Jerome Maire 2008-10
JM: nlens, w (initial guess), P (initial guess), cenx (or centrXpos), ceny (or centrYpos) as parameters
2009-09-17 JM: added DRF parameters
2009-12-10 JM: initiate position at 1.5microns so we can take into account several band
2010-07-14 JM:for DRP testing, correct for DST finite spectral resolution
2010-08-16 JM: added bad pixel map
2011-07-14 MP: Reworked FITS keyword handling to provide more informative
error messages in case of missing or invalid keywords.
2011-08-02 MP: Updated for multi-extension FITS.
2012-12-13 MP: Bad pixel map now taken from DQ extension if present.
Print more informative logging messages for the user
Various bits of code cleanup.
2012-12-20 JM: more centroid methods added
2013-07-12 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| nlens | int | [0,400] | 281 | side length of the lenslet array |
| centrXpos | int | [0,2048] | 1024 | Initial approximate x-position [pixel] of central peak at 1.5microns |
| centrYpos | int | [0,2048] | 1024 | Initial approximate y-position [pixel] of central peak at 1.5microns |
| w | float | [0.,10.] | 4.8 | Spectral spacing perpendicular to the dispersion axis at the image center [pixel] |
| P | float | [-7.,7.] | -1.8 | Ratio of spectral offset parallel to dispersion over spectral spacing perpendicular to dispersion |
| emissionlinesfile | string | None | AUTOMATIC | File of emission lines. |
| wav_of_centrXYpos | int | [1,2] | 2 | 1 if centrX-Ypos is the smallest-wavelength peak of the band; 2 if centrX-Ypos refer to 1.5microns |
| maxpos | float | [-7.,7.] | Allowed maximum location fluctuation (in pixel) between adjacent mlens | |
| maxtilt | float | [-360.,360.] | Allowed maximum tilt fluctuation (in degree) between adjacent mlens | |
| centroidmethod | int | [0,1] | 0 | Centroid method: 0 means barycentric (fast), 1 means gaussian fit (slow) |
| medfilter | int | [0,1] | 1 | 1: Median filtering of dispersion coeff and tilts with a (5x5) median filtering |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| iscalib | int | [0,1] | 1 | 1: save to Calibrations Database, 0: save in regular reduced data dir |
| lamp_override | int | [0,1] | 0 | 0,1: override the filter/lamp combinations? |
| gpitvim_dispgrid | int | [0,500] | 15 | 1-500: choose gpitv session for displaying image output and wavcal grid overplotted, 0: no display |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying wavcal file, 0: no display |
| tests | int | [0,3] | 0 | 1 for extensive tests |
| testsDST | int | [0,3] | 0 | 1 for DST tests |
IDL Filename: gpi_measure_wavelength_calibration.pro
Measure Polarization Spot Calibration¶
Derive polarization calibration files from a flat field image.
Category: Calibration Order: 1.8
Inputs: 2D image from flat field in polarization mode
Outputs: Measured polarization spot locations calibration file Output Suffix: Could not be determined automatically
Notes:
gpi_extract_polcal detects the positions of the polarized spots in a 2D
image based on flat field observations.
ALGORITHM:
gpi_extract_polcal starts by detecting the central peak of the image.
Next, starting with a initial value of w & P, it finds the nearest peak (with an increment on the microlens coordinates)
when nearest peak has been detected, it reevaluates w & P and so forth..
Like the spectral mode wavelength calibration code, the first part of this
algorithm is devoted to determining the positions of each spot on the
detector.
Unlike the spectral mode calibration, what we store here is in fact a
weighted list of pixels for each lenslet PSF. FIXME - this will need some
revision to accomodate flexure...
HISTORY:
2009-06-17: Started, based on gpi_extract_wavcal - Marshall Perrin
2009-09-17 JM: added DRF parameters
2013-01-28 MMB: added some keywords to pass to find_pol_positions_quadrant
2013-07-11 MDP: Documentation cleanup.
2013-07-12 MDP: Rename for consistency
2013-10-31 MMB: Big update to parallel code,
2014-11-25 MDP: Merge parallel and single threaded code into one primitive
with an option to switch between them.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| nlens | int | [0,400] | 281 | side length of the lenslet array |
| centrXpos | int | [0,2048] | 1078 | Initial approximate x-position [pixel] of central peak at 1.5microns |
| centrYpos | int | [0,2048] | 1028 | Initial approximate y-position [pixel] of central peak at 1.5microns |
| w | float | [0.,10.] | 4.4 | Spectral spacing perpendicular to the dispersion axis at the detcetor in pixel |
| P | float | [-7.,7.] | 2.18 | Micro-pupil pattern |
| maxpos | float | [-7.,7.] | 2.5 | Allowed maximum location fluctuation (in pixel) between adjacent mlens |
| FitWidth | float | [-10.,10.] | 3 | Size of box around a spot used to find center |
| parallel | Int | [0,1] | 0 | Option for Parallelization, 0: none, 1: parallel |
| Save | int | [0,1] | 1 | None |
| Display | int | [0,1] | 1 | None |
IDL Filename: gpi_measure_polarization_spot_calibration.pro
Assemble Polarization Cube¶
Extract 2 perpendicular polarizations from a 2D image.
Category: PolarimetricScience, Calibration Order: 2.0
Inputs: detector image in polarimetry mode
Outputs: Polarization pair datacube Output Suffix: ‘-podc’
Notes:
extract polarization-mode data cube from an image
define first suffix '-podc' (polarization data-cube)
This routine transforms a 2D detector image in the dataset.currframe input
structure into a 3D data cube in the dataset.currframe output structure.
(not much of a data cube - really just 2x 2D images)
ALGORITHM NOTES:
Ideally this should be done as an optimum weighting
(see e.g. Naylor et al, 1997 MNRAS)
That algorithm is as follows: For each lenslet spot,
-divide each pixel by the expected fraction of the total lenslet flux
in that pixel. (this makes each pixel an estimate of the total lenslet
flux)
-Combine these into a weighted average, weighted by the S/N per pixel
common needed: filter, wavcal, tilt, (nlens)
HISTORY:
2009-04-22 MDP: Created, based on DST's cubeextract_polarized.
2009-09-17 JM: added DRF parameters
2009-10-08 JM: add gpitv display
2010-10-19 JM: split HISTORY keyword if necessary
2011-07-15 MP: Code cleanup.
2011-06-07 JM: added FITS/MEF compatibility
2013-01-02 MP: Updated output file orientation to be consistent with
spectral mode and raw data.
2013-07-17 MP: Renamed for consistency
2013-11-30 MP: Clear DQ and Uncert pointers
2014-02-03 MP: Code and docs cleanup
2014-07-01 MPF: Modified "PSF" extraction for weighting by a Gaussian
and a noise map.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
| Method | String | BOX|PSF | BOX | Method for pol cube reconstruction, simple box or optimal PSF |
IDL Filename: gpi_assemble_polarization_cube.pro
Assemble Undispersed Image¶
Extract a 2D image from a raw undispersed mode image. Box integration of the light from each lenslet.
Category: Calibration Order: 2.0
Inputs: Not specified
Outputs:
Notes:
This routine performs a simple extraction of GPI IFS undispersed
data. It requires a pair of fits files explicitly named xlocs.fits
and ylocs.fits located in the current directory. Those files contain
a 300x300 array of x and y positions for spots. These files are
produced by the routine identify.pro which examines a flood illuminated
grid of spots.
The routine currently assumes the spots in the IFS are shifted by
2.36 and 2.63 pixels from the time the calibration frame was taken
in the UCLA lab. If your image has significant flux in the central lenslets
then you can comment out the fitting portion of the code, and the
pattern shift will be determined for you.
fname = name of the fits file you want to reduce
outname = name of the output file this routine will produce
example usage:
extu, "test0159.fits", "extu0159.fits"
KEYWORDS:
HISTORY:
Originally by James Larkin as extu.pro
2012-02-07 Pipelinified by Marshall Perrin
2012-03-30 Rotated by 90 deg to match spectral cube orientation. NaNs outside of FOV. - MP
2013-03-08 JM: added manual shifts of the spot due to flexure
2013-07-17 MP: Rename for consistency
2013-11-30 MDP: Clear DQ and Uncert pointers
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| xshift | float | [-100,100] | -2.363 | Shift in X direction |
| yshift | float | [-100,100] | -2.6134 | Shift in Y direction |
| boxsize | float | [0,10] | 5 | Size of box to use for spectral extraction |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| suffix | string | None | -extu | Enter output suffix |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_assemble_undispersed_image.pro
Assemble Spectral Datacube¶
Assemble a 3D datacube from a 2D image. Spatial integration (3 pixels box) along the dispersion axis
Category: SpectralScience, Calibration Order: 2.0
Inputs: Not specified
Outputs: Not specified Output Suffix: ‘-rawspdc’
Notes:
This routine transforms a 2D detector image in the dataset.currframe input
structure into a 3D data cube in the dataset.currframe output structure.
This routine extracts data cube from an image using spatial summation along the dispersion axis
introduced suffix '-rawspdc' (raw spectral data-cube)
KEYWORDS:
GEM/GPI KEYWORDS:IFSFILT
HISTORY:
Originally by Jerome Maire 2007-11
2012-02-01 JM: adapted to vertical dispersion
2012-02-15 JM: adapted as a pipeline module
2013-08-07 ds: idl2 compiler compatible
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| CalibrationFile | String | None | AUTOMATIC | Filename of the mlens-PSF calibration file to be read |
| ReuseOutput | int | [0,1] | 0 | 1: keep output for following primitives, 0: don’t keep |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_extractcube_mlenspsf.pro
Assemble Spectral Datacube using ePSF¶
Extract a 3D datacube from a 2D image. Spatial integration (3 pixels) along the dispersion axis
Category: SpectralScience, Calibration Order: 2.0
Inputs: Not specified
Outputs: Output Suffix: Could not be determined automatically
Notes:
This routine transforms a 2D detector image in the dataset.currframe input
structure into a 3D data cube in the dataset.currframe output structure.
This routine extracts data cube from an image using an inversion method along the dispersion axis
KEYWORDS:
GEM/GPI KEYWORDS:
HISTORY:
Originally by Jerome Maire 2007-11
2012-02-01 JM: adapted to vertical dispersion
2012-02-15 JM: adapted as a pipeline module
2013-08-07 ds: idl2 compiler compatible
2013-12-16 MP: CalibrationFile argument syntax update.
2014-07-18 JM: implemented ePSF instead of DST simulated PSF
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| CalibrationFile | String | None | AUTOMATIC | Filename of the mlens-PSF calibration file to be read |
| ReuseOutput | int | [0,1] | 0 | 1: keep output for following primitives, 0: don’t keep |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_extractcube_epsf.pro
Assemble Spectral Datacube¶
Assemble a 3D datacube from a 2D image. Spatial integration (3 pixels box) along the dispersion axis
Category: SpectralScience, Calibration Order: 2.0
Inputs: Not specified
Outputs: Not specified Output Suffix: ‘-rawspdc’
Notes:
This routine transforms a 2D detector image in the dataset.currframe input
structure into a 3D data cube in the dataset.currframe output structure.
This routine extracts data cube from an image using spatial summation along the dispersion axis
introduced suffix '-rawspdc' (raw spectral data-cube)
KEYWORDS:
GEM/GPI KEYWORDS:IFSFILT
HISTORY:
Originally by Jerome Maire 2007-11
2008-04-02 JM: spatial summation window centered on pixel and interpolation on the zem. comm. wav. vector
2008-06-06 JM: adapted to pipeline inputs
2009-04-15 MDP: Documentation updated.
2009-06-20 JM: adapted to wavcal input
2009-09-17 JM: added DRF parameters
2012-02-01 JM: adapted to vertical dispersion
2012-02-09 DS: offloaded sdpx calculation
2013-04-02 JBR: Correction on the y coordinate when reading the det array to match centered pixel convention. Removal of the reference pixel area.
2013-07-17 MDP: Rename for consistency
2013-08-06 MDP: Documentation update, code cleanup to relabel X and Y properly
2013-11-30 MDP: Clear DQ and Uncert pointers
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_assemble_spectral_datacube.pro
Filter datacube spatially¶
Apply spatial filter to datacubes
Category: ALL Order: 2.1
Inputs: raw 2D image file
Outputs: 2D image corrected for dark current Output Suffix: ‘sfilt’
Notes:
High-pass or Low-pass filter each slice of a GPI datacube using a median box filter.
This is useful for removing the halos created by uncorrected atmospheric turbulence. This is a tad slow but a useful tool.
Skip_parallelization should be used when opening IDL sessions is slow (when using floating license servers etc)
Other filters may be added later.
HISTORY:
Originally by Patrick Ingraham Apr 2, 2014
MMB updated to include lowpass and polarimetry.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| hp_boxsize | int | [0,50] | 0 | 0: no filter, 1+: Filter box size |
| high_or_lowpass | string | [high|low] | high | High pass or lowpass filter? |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_filter_datacube_spatially.pro
Noise and Flux Analysis¶
Store a few key values as fits keywords in the file. It can generate anciliary files too.
Category: HIDDEN Order: 2.1
Inputs: Not specified
Outputs: Changes is the header of the file without changing the data and saving a fits file report with the value of the sliding median/standard deviation computation. Output Suffix: Could not be determined automatically
Notes:
/!\ HIDDEN /!\ It was a primitive used by JB for debug but I don't think it is gonna be used by anyone else.
This routine quantifies the noise and the flux in an image without changing it. It generates fits keyword for this values for further easy image sorting.
If asked, it can generate a fits files too.
If Flux = 1: Generate fits keywords related with total flux in the image
DN, total data number of the image.
DNLENS, total data number in the lenslets aera (if not a dark and not a cube)
DNBACK, total data number outside the lenslets aera (if not a dark and not a cube)
If StddevMed > 1: Generate fits keywords related with the standard deviation in the image
If StddevMed = 2:
Compute the local median and the local standard deviation by moving a square of size Width.
Because it is time consuming, you can skip pixels using the parameter PixelsSkipped.
In the output, the finite value pixels correspond to pixels where the media and the standard deviation were computed.
If 2d image: Generate a file with the suffix '-stddevmed' containing an 3d array. [*,*,0] is the median and [*,*,1] is the standard deviation.
If 3d image: Generate two files '-stddev' and '-median'. Both same size of the original image.
If microNoise = 1:
Estimate the quantity of microphonics noise in the image based on a model stored as a calibration file.
The quantity of microphonics noise is measured with the ratio of the dot_product and the norm of the image: dot_product/sqrt(sum(abs(fft(image))^2)).
With dot_product = sum(abs(fft(image))*abs(fft(noise_model))) which correspond to the projection of the image on the microphonics noise model in the absolute Fourier space.
The fits keyword associated is MICRONOI.
If FourierTransf = 1 or 2:
Build and save the Fourier transform of the image.
If 1, the output is the one directly from the idl function (fft). Therefore, the Fourier image is not centered.
If 2, the output will be centered.
In the case of a cube it is not a 3d fft that is performed but several 2d ffts.
suffix='-absfft' or suffix='-absfftdc' if it is a cube.
HISTORY:
Originally by Jean-Baptiste Ruffio 2013-05
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Flux | int | [0,1] | 1 | Trigger flux analysis |
| StddevMed | int | [0,2] | 1 | Trigger the standard deviation (and median) analysis of the image. if StddevMed=1, only keywords and log are produced. If StddevMed=2, fits files are generated with a sliding median and standard deviation. |
| Width | int | [3,2048] | 101 | If Stddev = 2, Width of the moving rectangle. It has to be odd. |
| PixelsSkipped | int | [0,2047] | 100 | If Stddev = 2, Pixels skipped between two points |
| MicroNoise | int | [0,1] | 1 | Trigger the microphonics noise analysis |
| CalibrationFile | string | None | AUTOMATIC | Filename of the desired microphonics model file to be read |
| FourierTransf | int | [0,1,2] | 1 | 1: frequency 0 on the bottom left. 2: frequencies 0 will be centered on the image. |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_noise_and_flux_analysis.pro
Divide by Lenslet Flat Field¶
Divides a spectral data-cube by a flat field data-cube.
Category: SpectralScience,PolarimetricScience,Calibration Order: 2.2
Inputs: Spectral or polarization datacube
Outputs: Each slice of the input datacube is divided by the lenslet flat.
Notes:
HISTORY:
2014-01-02 MP: New primitive
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_divide_by_lenslet_flat_field.pro
Divide by Spectral Flat Field¶
Divides a spectral data-cube by a flat field data-cube.
Category: SpectralScience,Calibration Order: 2.2
Inputs: data-cube
Outputs: Flat fielded datacube
Notes:
** Needs additional work, will not produce high qualty results yet **
HISTORY:
2009-08-27: JM created
2009-09-17 JM: added DRF parameters
2009-10-09 JM added gpitv display
2010-10-19 JM: split HISTORY keyword if necessary
2011-07 JM: added check for NAN & zero
2012-10-11 MP: Added min/max wavelength checks
2012-10-17 MP: Removed deprecated suffix= keyword
2013-07-17 MP: Rename for consistency
2013-12-30 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_divide_by_spectral_flat_field.pro
Remove Flat Lamp spectrum¶
Fit the lamp spectrum and remove it (for delivering flat field cubes)
Category: Calibration Order: 2.25
Inputs: Flat field data-cube
Outputs: Flat datacube normalized to remove lamp spectrum Output Suffix: ‘specflat’
Notes:
Rescale flat-field (keep large scale variations)
**CAUTION needs additional improvement **
GEM/GPI KEYWORDS:
DRP KEYWORDS: FILETYPE, ISCALIB
HISTORY:
2009-06-20 JM: created
2009-07-22 MP: added doc header keywords
2012-10-11 MP: added min/max wavelength checks
2013-07-17 MP: Rename for consistency
2013-12-03 MP: Add check for GCALLAMP=QH on input images
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
| method | string | polyfit|linfit|blackbody|none | blackbody | Method to use for removing lamp spectrum |
IDL Filename: gpi_remove_flat_lamp_spectrum.pro
Interpolate Wavelength Axis¶
Interpolate spectral datacube onto regular wavelength sampling.
Category: SpectralScience,Calibration Order: 2.3
Inputs: A raw irregularly-sampled spectral datacube
Outputs: Spectral datacube with slices at a regular wavelength sampling Output Suffix: ‘spdc’
Notes:
Interpolate datacube to have each slice at the same wavelength.
This is a necessary step of creating datacubes in spectral mode
and should always be used right after Assemble Spectral Datacube.
Also adds wavelength keywords to the FITS header.
HISTORY:
Originally by Jerome Maire 2008-06
2009-04-15 MDP: Documentation improved.
2009-06-20 JM: adapted to wavcal
2009-09-17 JM: added DRF parameters
2010-03-15 JM: added error handling
2012-12-09 MP: Updates to WCS output
2013-07-12 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Spectralchannels | int | [0,100] | 37 | Choose how many spectral channels for output datacube |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_interpolate_wavelength_axis.pro
Subtract Thermal/Sky Background Cube if K band¶
Subtract a thermal/sky cube
Category: ALL Order: 2.35
Inputs: 3D image file
Outputs: 3D image file, unchanged if YJH, background subtracted if K1 or K2. Output Suffix: ‘bkgnd_cube_sub’
Notes:
Subtract thermal background emission in the datacube, for K band data only
This is identical to the gpi_subtact_thermal_sky_if_k_band primtive except the subtraction
is done in cube space instead of detector space. It also uses sky cubes rather than the 2d sky images.
** special note: **
This is a new kind of "data dependent optional primitive". If the filter of
the current data is YJH, return without doing *anything*, even logging the
start/end of this primitive. It becomes a complete no-op for non-K-band
cases.
Algorithm:
Get the best available thermal/sky background cube calibration file from CalDB
Scale it to current exposure time
Subtract it.
The name of the calibration file used is saved to the DRPBKGND header keyword.
ALGORITHM TODO: Deal with uncertainty and pixel mask frames too.
HISTORY:
2013-12-23 PI: Initial implementation
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Name of thermal/sky background cube to subtract |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| Override_scaling | float | [0,10] | 1.0 | Set to value other than 1 to manually adjust the background image flux scaling to better match the science data |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_subtract_thermal_sky_background_cube_if_k_band.pro
Check for closed-loop coronagraphic image¶
Check whether file represents a closed-loop coronagraphic image.
Category: Calibration,SpectralScience,PolarimetricScience Order: 2.41
Inputs: Not specified
Outputs: Not specified
Notes:
This primitive checks that the input file is in fact a coronagraphic image.
It is intended to be used in quicklook recipes that may encounter all sorts
of different data.
Any following primitives will only be executed if the
image is in fact coronagraphic data. This is useful so the quicklook
recipe can include satellite spots or contrast measurement primitives,
which would generally cause the recipe to fail if they receive any
unocculted data. With this primitive added in the recipe before those
steps, they will just be skipped without producing any error messages.
HISTORY:
2013-08-02 ds - initial version
2013-11-12 MP - add check for PUPVIEWR inserted
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| err_on_false | int | [0,1] | 0 | If false, 0: continue to next image; 1: Throw error |
IDL Filename: gpi_check_coronagraph_status.pro
Measure GPI distortion from grid pattern¶
Measure GPI distortion from grid pattern
Category: Calibration HIDDEN Order: 2.44
Inputs: Not used
Outputs: Distortion correction coefficients file (values are currently hard-coded) Output Suffix: ‘-distor’
Notes:
CAUTION - NOT IMPLEMENTED
Note that this primitive can go *either* before or after
Accumulate Images.
As a Level 1 primitive, it will undistort one cube at a time;
As a Level 2 primitive it will undistort the whole stack of
accumulated images all at once.
This primitive *MUST* be run before the 'Measure Satellite spot locations' primitive
HISTORY:
Originally by Jerome Maire 2009-12
Switched sxaddpar to backbone->set_keyword 01.31.2012 Dmitry Savransky
2013 Added hard-coded measurement of the distortion made by Quinn JM
2014-05-10 MP: Update to enable this to work before or after accumulate
images.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
IDL Filename: gpi_measure_distortion.pro
Measure satellite spot locations¶
Measure the locations of the satellite spots in the datacube, and save the results to the FITS keyword headers.
Category: Calibration,SpectralScience Order: 2.44
Inputs: Not specified
Outputs: Not specified
Notes:
Measures the locations of the satellite spots; saves to FITS keywords.
The sat spots locations are saved to SATS1_1, SATS1_2, and so on.
The inferred location of the star is saved to PSFCENTX and PSFCENTY
(this is the mean location of all the locations at each wavelength)
By default, the sat spots information are saved to the FITS header keywords
of the current file in memory, and will only be saved if you subsequently
save that datacube (i.e. using 'save=1' on this primitive or a subsequent
one). The 'update_prev_fits_header' option will, in addition, also let you
write the same keyword information to the header of the most recently saved
file. This is useful if you have just already saved the datacube, and you
only now want to update this metadata.
HISTORY:
Originally by Jerome Maire 2009-12
2012-09-18 Offloaded functionality to common backend - ds
2013-07-17 MP Documentation updated, rename for consistency.
2014-10-11 DS Added PSFCEN_XX for all slices to header
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| refine_fits | int | [0,1] | 1 | 0: Use wavelength scaling only; 1: Fit each slice |
| reference_index | int | [-1,50] | -1 | Index of slice to use for initial satellite detection. -1 for Auto. |
| search_window | int | [1,50] | 20 | Radius of aperture used for locating satellite spots. |
| highpass | int | [0,25] | 1 | 1: Use high pass filter (default size) 0: don’t 2+: size of highpass filter box |
| constrain | int | [0,1] | 0 | 1: Constrain distance between sat spots by band; 0: Unconstrained search. |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| update_prev_fits_header | int | [0,1] | 0 | Update FITS metadata in the most recently saved datacube? |
| loc_input | int | [0,2] | 0 | 0: Find spots automatically; 1: Use values below as initial satellite spot location |
| x1 | int | [0,300] | 0 | approx x-location of top left spot on reference slice of the datacube in pixels |
| y1 | int | [0,300] | 0 | approx y-location of top left spot on reference slice of the datacube in pixels |
| x2 | int | [0,300] | 0 | approx x-location of bottom left spot on reference slice of the datacube in pixels |
| y2 | int | [0,300] | 0 | approx y-location of bottom left spot on reference slice of the datacube in pixels |
| x3 | int | [0,300] | 0 | approx x-location of top right spot on reference slice of the datacube in pixels |
| y3 | int | [0,300] | 0 | approx y-location of top right spot on reference slice of the datacube in pixels |
| x4 | int | [0,300] | 0 | approx x-location of bottom right spot on reference slice of the datacube in pixels |
| y4 | int | [0,300] | 0 | approx y-location of bottom right spot on reference slice of the datacube in pixels |
IDL Filename: gpi_measure_satellite_spot_locations.pro
Measure GPI distortion from grid pattern¶
Measure GPI distortion from grid pattern
Category: Calibration HIDDEN Order: 2.44
Inputs: Not used
Outputs: Distortion correction coefficients file (values are currently hard-coded) Output Suffix: ‘-distor’
Notes:
CAUTION - NOT IMPLEMENTED
the distortion in lab was analyzed with other non-pipeline tools by Quinn,
and the result is just hard coded here to output it in the GPI pipeline
format.
HISTORY:
Originally by Jerome Maire 2009-12
Switched sxaddpar to backbone->set_keyword 01.31.2012 Dmitry Savransky
2013 Added hard-coded measurement of the distortion made by Quinn JM
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
IDL Filename: gpi_measure_distortion.pro
Measure Star Position for Polarimetry¶
Finds the location of the occulted star in polarimetry mode, and save the results to the FITS keyword headers.
Category: Calibration, PolarimetricScience Order: 2.445
Inputs: Polarimetric mode datacube
Outputs: Polarimetric mode datacube with star location recorded in header
Notes:
Finds the location of the occulted star (i.e. image center); saves center to FITS keywords.
The algorithm used is a version of the radon transform, used to find where
all the broadband extended speckles intersect in the image.
The inferred star location is saved to PSFCENTX, PSFCENY keywords in the
header
HISTORY:
2014-01-31 JW: Created. Accurary is subpixel - hopefully.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| x0 | int | [0,300] | 147 | initial guess for image center x-coordinate |
| y0 | int | [0,300] | 147 | inital guess ofr image center y-coordinate |
| search_window | int | [1,50] | 5 | Radius of search window to search for the center |
| mask_radius | int | [0,100] | 50 | Radius of center of image to mask (centered on x0, y0 inputs) |
| highpass | int | [0,1] | 1 | 1: Use high pass filter 0: don’t |
| lower_threshold | float | [-100000,100000] | -100 | Lower pixel values will be converted to this value |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_measure_star_position_for_polarimetry.pro
Measure satellite spot peak fluxes¶
Calculate peak fluxes of satellite spots in datacubes
Category: Calibration,SpectralScience Order: 2.45
Inputs: spectral datacube with spot locations in the header
Outputs: datacube with measured spot fluxes
Notes:
Measure the fluxes of the satellite spots.
You must run 'Measure Satellite Spot Locations' before you can use this
one.
Spot fluxes are measured and then saved to SATF1_1, SATF1_2 etc keywords
in the header.
By default, the sat spots information are saved to the FITS header keywords
of the current file in memory, and will only be saved if you subsequently
save that datacube (i.e. using 'save=1' on this primitive or a subsequent
one). The 'update_prev_fits_header' option will, in addition, also let you
write the same keyword information to the header of the most recently saved
file. This is useful if you have just already saved the datacube, and you
only now want to update this metadata.
HISTORY:
Written 09-18-2012 savransky1@llnl.gov
2013-07-17 MP: Renamed for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| gauss_fit | int | [0,1] | 1 | 0: Extract maximum pixel; 1: Correlate with Gaussian to find peak |
| reference_index | int | [0,50] | 0 | Index of slice to use for initial satellite detection. |
| ap_rad | int | [1,50] | 7 | Radius of aperture used for finding peaks. |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| update_prev_fits_header | int | [0,1] | 0 | Update FITS metadata in the most recently saved datacube? |
IDL Filename: gpi_measure_satellite_spot_peak_fluxes.pro
Correct for Atmospheric Differential Refraction¶
Interpolates the cube to undo any shifts due to ADR (or leftover ADC).
Category: Calibration, SpectralScience Order: 2.45
Inputs: Spectral Cube
Outputs: Interpolated Spectral Cube that is shifted for ADR compensation
Notes:
Uses the location of the satellite spots to calculate a center at each
wavelength, fits the x and y drift as a function of wavelenght to a straight
line and compensates for it.
HISTORY:
2014-01-31 JW: Created. Accurary is subpixel - hopefully.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| refslice | int | [0,36] | 20 | reference slice to perform relative shifts |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_correct_adr_shift.pro
Extract one spectrum, plots¶
Extract one spectrum from a datacube somewhere in the FOV specified by the user.
Category: SpectralScience Order: 2.51
Inputs: data-cube
Outputs: datacube calibrated for telluric absorption
Notes:
Corrects for telluric transmission based on a telluric calibrator
target reduced and provied as the CalibrationFile.
HISTORY:
2009-08-27: JM created
2009-09-17 JM: added DRF parameters
2009-10-09 JM added gpitv display
2010-10-19 JM: split HISTORY keyword if necessary
2011-08-01 MP: Update for multi-extension FITS files
2012-10-10 MP: Minor code cleanup; remove deprecated suffix= parameter
2013-07-17 MP: Rename for consistency
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_divide_by_telluric_transmission.pro
Interpolate bad pixels in cube¶
Repair bad pixels by interpolating between their neighbors.
Category: SpectralScience, PolarimetricScience, Calibration Order: 2.5
Inputs: Cube in either spectral or polarization mode
Outputs: Cube with bad pixels potentially found and cleaned up. Output Suffix: ‘-bpfix’
Notes:
Searches for statistical outlier bad pixels in a cube and replace them
by interpolating between their neighbors.
CAUTION:
Heuristic and not guaranteed or tested in any way; this is more a
convenience function than a rigorous statistcally justified repair tool
HISTORY:
2013-12-14 MP: Created as a convenience function cleanup tool. Almost
certainly not the best algorithm - just something quick and good enough for
now?
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 1 | 1-500: choose gpitv session for displaying output, 0: no display |
| before_and_after | int | [0,1] | 0 | Show the before-and-after images for the user to see? (for debugging/testing) |
IDL Filename: gpi_interpolate_bad_pixels_in_cube.pro
Extract one spectrum¶
Extract one spectrum from a datacube somewhere in the FOV specified by the user.
Category: SpectralScience HIDDEN Order: 2.51
Inputs: data-cube
Outputs: Output Suffix: Could not be determined automatically
Notes:
KEYWORDS:
/Save Set to 1 to save the output image to a disk file.
GEM/GPI KEYWORDS:FILTER,IFSUNIT
DRP KEYWORDS: CUNIT,DATAFILE,SPECCENX,SPECCENY
HISTORY:
JM 2010-03 : created module.
2012-10-17 MP: Removed deprecated suffix keyword. FIXME this needs major
cleanup!
2013-08-07 ds: idl2 compiler compatible
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| xcenter | int | [0,1000] | 141 | x-locations in pixel on datacube where extraction will be made |
| ycenter | int | [0,1000] | 141 | y-locations in pixel on datacube where extraction will be made |
| radius | float | [0,1000] | Aperture radius (in pixel i.e. mlens) to extract photometry for each wavelength. | |
| method | string | [median|mean|total] | total | method of photometry extraction:median,mean,total |
| ps_figure | int | [0,500] | 2 | 1-500: choose # of saved fig suffix name, 0: no ps figure |
| Save | int | [0,1] | 1 | 1: save output (fits) on disk, 0: don’t save |
IDL Filename: extract_one_spectrum.pro
Calibrate Photometric Flux¶
Apply photometric calibration to a single or set of datacubes
Category: SpectralScience Order: 2.51
Inputs:
Outputs: Not specified Output Suffix: ‘-phot’
Notes:
This primitive applies a spectrophotometric calibrations to the datacube that is determined either
from the satellite spots of the supplied cube, the satellite spots of a
user-indicated cube, or any user-supplied spectral response function (e.g. derived
from an open loop image of a standard star). Use of this primitive is complicated, therefore a
tutorial has been written to guide the user through it's use. The tutorial can be found as part
of the online documentation under the User's guide (http://docs.planetimager.org/pipeline/usage/index.html).
The user may also specify the extraction and sky radii used in performing the aperture photometry. Note that the 'annuli' only represent the radial size of the extraction. The background is extracted by fitting a constant to an annulus surrounding the central star at the same radius as the planet. The inner width of the annulus is equal to the inner_sky_radius, the outer annulus describes the distance from the companion to the edges of the annulus that should be considered when fitting the constant. If the user wishes to examine the section being fit, they should modify line 350 accordingly.
Error bars are calculated and put into the headers to be used with future primitives such as gpi_extract_1d_spectrum. They are determined by convolving the sky annulus with the extraction aperture then taking the standard deviation.
1: (REQUIRED) datacube that requires calibration (loaded as an Input FITS file). The satellites of this image are used to determine the spectrophotometry calibration unless one of the options below are used.
2a (OPTIONAL): datacube or to be used to determine the calibration - this should be entered into the calib_cube_name parameter. This is meant to be used when the satellites of a single cube are too low SNR and a stack of cubes must be used to get the SNR to high enough levels (which naturally blurs the companion).
OR
2b (OPTIONAL): a 2D spectrum (in ADU per COADD, where the COADD corresponds to input #1) - this should be entered using the calib_spectrum keyword. The file format must be three columns, the first being wavelength in microns, the second being flux in ADU per COADD, and third being the uncertainty. THis is useful to calibrate a cube if the calibration uses a different star as a calibrator.
The spectral type of the star and it's magnitude must be defined in the SPECTYPE and HMAG header keywords. These should be set by default but sometimes are not (or are not set correctly). One can modify them using the <Add Gemini and GPI keywords> primitive. Note that this primitive is only visible when <Show all primitives> is selected from the <Options> dropdown menu in the recipe editor. Based on these values, the primitive searches for an appropropriate pickles model for the spectral type and uses this to perform the calibration. If the pickles model is not appropriate, or you wish to provide a different model, this model can be input using the 'calib_model_spectrum' keyword. The file format must be three columns, the first being wavelength in microns, the second being the flux in erg/s/cm2/A, the third being the uncertainty.
Note that the calib_cube_name,calib_spectrum, and calib_model_spectrum require the entire directory+filename unless they are in the output directory
GEM/GPI KEYWORDS:FILTER,IFSUNIT
DRP KEYWORDS: CUNIT,DATAFILE
HISTORY:
JM 2010-03 : created module.
2012-10-17 MP: Removed deprecated suffix keyword. needs major cleanup!
2013-08-07 ds: idl2 compiler compatible
2014-01-07 PI: Created new gpi_calibrate_photometric_flux - big overhaul from the original apply_photometric_calibration
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
| extraction_radius | float | [0,1000] | Aperture radius at middle wavelength (in spaxels i.e. mlens) to extract photometry for each wavelength. | |
| inner_sky_radius | float | [1,100] | Inner aperture radius at middle wavelength (in spaxels i.e. mlens) to extract sky for each wavelength. | |
| outer_sky_radius | float | [1,100] | Outer aperture radius at middle wavelength (in spaxels i.e. mlens) to extract sky for each wavelength. | |
| c_ap_scaling | int | [0,1] | 1 | Perform aperture scaling with wavelength? |
| calib_cube_name | string | None | Leave blank to use satellites of this cube, or enter a file to use those satellites | |
| calib_model_spectrum | string | None | Leave blank to use satellites of this cube, or enter a file to use with the spectrum for the satellites | |
| calib_spectrum | string | None | Leave blank to use satellites of this cube, or enter calibrated spectrum file | |
| FinalUnits | int | [0,10] | 5 | 0: ADU per coadd, 1: ADU/s, 2: ph/s/nm/m^2, 3: Jy, 4: ‘W/m^2/um, 5: ergs/s/cm^2/A, 6: ergs/s/cm^2/Hz’ |
IDL Filename: gpi_calibrate_photometric_flux.pro
Extract one spectrum, plots¶
Extract one spectrum from a datacube somewhere in the FOV specified by the user.
Category: SpectralScience HIDDEN Order: 2.51
Inputs: data-cube
Outputs: Output Suffix: Could not be determined automatically
Notes:
KEYWORDS:
/Save Set to 1 to save the output image to a disk file.
KEYWORDS:
GEM/GPI KEYWORDS:FILTER,IFSUNIT
DRP KEYWORDS: CUNIT,DATAFILE,SPECCENX,SPECCENY
HISTORY:
JM 2010-03 : created module.
2013-08-07 ds: idl2 compiler compatible
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| xcenter | int | [0,1000] | 141 | x-locations in pixel on datacube where extraction will be made |
| ycenter | int | [0,1000] | 141 | y-locations in pixel on datacube where extraction will be made |
| radius | float | [0,1000] | Aperture radius (in pixel i.e. mlens) to extract photometry for each wavelength. | |
| method | string | [median|mean|total] | total | method of photometry extraction:median,mean,total |
| ps_figure | int | [0,500] | 2 | 1-500: choose # of saved fig suffix name, 0: no ps figure |
| Save | int | [0,1] | 1 | 1: save output (fits) on disk, 0: don’t save |
| suffix | string | None | -spec | Enter output suffix (fits) |
IDL Filename: extract_one_spectrum2.pro
Extract 1D spectrum from a datacube¶
Extract one spectrum from a datacube somewhere in the FOV specified by the user.
Category: SpectralScience Order: 2.52
Inputs: Datacube containing a source that needs extracting, located by the xcenter and ycenter arguments
Outputs: 1D spectrum Output Suffix: Could not be determined automatically
Notes:
WARNING: This primitive will not provide spectra of publishable quality
it is designed to perform a quick extraction of a source.
This primitive extracts a spectrum from a data cube. It is meant to be used
on datacubes that have been calibrated by gpi_apply_photometric_calibration,
but this is not strictly required.
The extraction radius is pulled out of the header such that is uses the
same as what was used to calibrate the cube. If they keyword is not found,
then the extraction_radius keyword is used. The extraction_radius keyword will
also be used if the override keyword is set to 1. Note that this is NOT
recommended and will introduce systematics into the data.
The centroiding is performed by fitting a gaussian to the region of interest.
A line is then fit to the centroids and used. In this fit, the first and last
4 data points are excluded due to low transmission. The errors for each centroid
are determined by taking the largest of the offsets between the subtraction of adjacent
centroids (e.g. yerr[j]=0.1>abs(yarr0[j]-yarr0[j+1])>abs(yarr0[j]-yarr0[j-1]) )
All photometry is done in ADU/coadd. This is performed by converting the cube to
ADU/coadd then converting back to whatever units the cube was input with.
The error bars are determined using the same method as the satellite spots
in gpi_calibrate_photometric_flux primitive. The user specifies the sky radii used in performing the aperture photometry. Note that the 'annuli' only represent the radial size of the extraction. The background is extracted by fitting a constant to an annulus surrounding the central star at the same radius as the planet. The inner width of the annulus is equal to the inner_sky_radius, the outer annulus describes the distance from the companion to the edges of the annulus that should be considered when fitting the constant. If the user wishes to examine the section being fit, they should modify line 350 accordingly.
Highpass filtering the image is recommended to determine the centroids, note that the highpass filtered image
is not used when measuring the extracted spectrum.
KEYWORDS:
Save: Set to 1 to save the spectrum to a disk file (.fits).
xcenter: x-location of extraction (in pixels)
ycenter: y-location of extraction (in pixels)
highpass: highpass filter the image when determining centroid?
inner_sky_radius: inner radius used in defining sky subtraction annulus section
outer_sky_radius: outer radius used in defining sky subtraction annulus section
override: allows input of a new extraction radius, and the use/non-use of c_ap_scaling
extraction_radius: Radius used to define annulus for source extraction. This keyword is only active if the override keyword is set, or if the CEXTR_AP keyword, set by the Calibrate Photometric Flux primitive (gpi_calibrate_photometric_flux.pro) is not present in the header
c_ap_scaling: keyword that activates the scaling of the apertures with wavelength. This keyword is only active if the override keyword is set, or if the C_AP_SC keyword, set by the Calibrate Photometric Flux primitive (gpi_calibrate_photometric_flux.pro) is not present in the header
display: window used to display the extracted spectrum plot
save_ps_plot: saves a postscript version of the plot if desired
write_ascii_file: writes as ascii output of the spectra - no header info included
GEM/GPI KEYWORDS:FILTER,IFSUNIT
DRP KEYWORDS: CUNIT,DATAFILE,SPECCENX,SPECCENY
HISTORY:
2014-01-07 PI: Created Module - big overhaul from the original extract 1d spectrum
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| xcenter | float | [-1,280] | -1 | x-location in pixels on datacube where extraction will be made |
| ycenter | float | [-1,280] | -1 | y-location in pixels on datacube where extraction will be made |
| highpass | int | [0,25] | 0 | highpass filter box size for centroiding |
| no_centroid_override | int | [0,1] | 0 | Do not centroid on extraction source? |
| inner_sky_radius | float | [1,100] | Inner aperture radius at middle wavelength slice (in spaxels i.e. mlens) to extract sky | |
| outer_sky_radius | float | [1,100] | Outer aperture radius at middle wavelength slice (in spaxels i.e. mlens) to extract sky | |
| override | int | [0,1] | 0 | Override apertures/scaling from spectrophotometric calibration? |
| extraction_radius | float | [0,1000] | Aperture radius at middle wavelength (in spaxels i.e. mlens) to extract photometry for each wavelength. (only active if Override is set) | |
| c_ap_scaling | int | [0,1] | 1 | Perform aperture scaling with wavelength? |
| display | int | [-1,100] | 17 | -1 = No display; 0 = New (unused) window; else = Window number to display diagnostic plot. |
| save_ps_plot | int | [0,1] | 0 | Save PostScript of plot? |
| write_ascii_file | int | [0,1] | 0 | Save ascii file of spectrum (.dat)? |
IDL Filename: gpi_extract_1d_spectrum.pro
Collapse datacube¶
Collapse the wavelength dimension of a datacube via mean, median or total.
Category: ALL Order: 2.6
Inputs: any datacube
Outputs: image containing that datacube collapsed to 2D Output Suffix: Could not be determined automatically
Notes:
TODO: more advanced collapse methods.
GEM/GPI KEYWORDS:
DRP KEYWORDS: CDELT3, CRPIX3,CRVAL3,CTYPE3,NAXIS3
HISTORY:
2010-04-23 JM created
2011-07-30 MP: Updated for multi-extension FITS
2013-07-12 MP: primitive rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | enum | MEDIAN|TOTAL | TOTAL | How to collapse datacube: total or median (with flux conservation) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| ReuseOutput | int | [0,1] | 1 | 1: keep output for following primitives, 0: don’t keep |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_collapse_datacube.pro
Simple Spectral Differential Imaging¶
Apply SSDI to create a 2D subtracted image from a cube. Given the user’s specified wavelength ranges, extract the 3D datacube slices for each of those wavelength ranges. Collapse these down into 2D images by simply averaging the values at each lenslet (ignoring NANs). Rescale Image1, then compute diffImage = I1scaled - k* I2
Category: SpectralScience Order: 2.61
Inputs:
Outputs: Output Suffix: Could not be determined automatically
Notes:
This recipe rescales and subtracts 2 frames in different user-defined
bandwidths. This recipe is used for speckle suppression using the
Marois et al (2000) algorithm.
This routine does NOT update the data structures in memory. You **MUST**
set the keyword SAVE=1 or else the output is silently discarded.
input datacube
wavelength solution from common block
KEYWORDS:
L1Min= Wavelength range 1, minimum wavelength [in microns]
L1Max= Wavelength range 1, maximum wavelength [in microns]
L2Min= Wavelength range 2, minimum wavelength [in microns]
L2Max= Wavelength range 2, maximum wavelength [in microns]
k= Multiplicative coefficient for multiplying the image for
Wavelength Range *2*. Default value is k=1.
/Save Set to 1 to save the output file to disk
DRP KEYWORDS: CDELT3,CRPIX3,CRVAL3,CTYPE3,NAXIS3
ALGORITHM:
Given the user's specified wavelength ranges, extract the 3D datacube slices
for each of those wavelength ranges. Collapse these down into 2D images by
simply averaging the values at each lenslet (ignoring NANs). Rescale Image1
using fftscale so that the PSF scale matches that of Image2 (as computed
from the average wavelength for each image). Then compute
diffImage = I1scaled - k* I2
Then hopefully output the image somewhere if SAVE=1 is set.
HISTORY:
2007-11 Jerome Maire
2009-04-15 MDP: Documentation updated; slight code cleanup
2009-09-17 JM: added DRF parameters
2013-08-07 ds: idl2 compiler compatible
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| L1Min | float | [0.9,2.5] | 1.55 | Wavelength range 1, minimum wavelength [in microns] |
| L1Max | float | [0.9,2.5] | 1.57 | Wavelength range 1, maximum wavelength [in microns] |
| L2Min | float | [0.9,2.5] | 1.60 | Wavelength range 2, minimum wavelength [in microns] |
| L2Max | float | [0.9,2.5] | 1.65 | Wavelength range 2, maximum wavelength [in microns] |
| k | float | [0,10] | 1.0 | Scaling factor of Intensity(wav_range2) with diffImage = I1scaled - k* I2 |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| ReuseOutput | int | [0,1] | 1 | 1: keep output for following primitives, 0: don’t keep |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_simple_spectral_differential_imaging.pro
Calibrate astrometry from binary (using 6th orbit catalog)¶
Calculate astrometry from unocculted binaries; Calculate Separation and PA at date DATEOBS using the sixth orbit catalog.
Category: Calibration Order: 2.61
Inputs: spectral mode datacube, observing a reference binary
Outputs: plate scale & orientation, saved as calibration file Output Suffix: ‘astrom’ ; output suffix
Notes:
GEM/GPI KEYWORDS:CRPA,DATE-OBS,OBJECT,TIME-OBS
DRP KEYWORDS: FILETYPE,ISCALIB
HISTORY:
Originally by Jerome Maire 2009-12
2013-07-19 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
IDL Filename: gpi_calibrate_astrometry_from_binary_orbit_catalog.pro
Speckle alignment¶
This recipe rescales datacube PSF slices with respect to a chosen reference PSF slice.
Category: SpectralScience Order: 2.61
Inputs: Spectral datacube
Outputs: Resampled and rescaled spectral datacube Output Suffix: suffix+’-specalign’
Notes:
This recipe rescales datacube slices with respect to a chosen reference slice.
ALGORITHM:
Given the user's specified wavelength ranges, extract the 3D datacube slices
for each of those wavelength ranges. Rescale slices with respect to a reference slice
using fftscale so that the PSF scale matches that of reference (as computed
from the average wavelength for each image).
HISTORY:
2012-02 JM
07.30.2012 - offladed backend to speckle_align - ds
05.10.2012 - updates to match other primitives and to reflect
changes to backend by Tyler Barker
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| k | int | [0,100] | 0 | Slice of the reference PSF |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_speckle_alignment.pro
Measure Contrast¶
Measure the contrast. Save as PNG or FITS table.
Category: SpectralScience,PolarimetricScience Order: 2.7
Inputs: Spectral mode datacube
Outputs: Contrast datacube, plot of contrast curve Output Suffix: ‘-contr’
Notes:
Measure, display on screen, and optionally save the contrast.
TODO - should we revise this to call the same contrast measurement backend
as GPItv?
By default, the sat spots information are saved to the FITS header keywords
of the current file in memory, and will only be saved if you subsequently
save that datacube (i.e. using 'save=1' on this primitive or a subsequent
one). The 'update_prev_fits_header' option will, in addition, also let you
write the same keyword information to the header of the most recently saved
file. This is useful if you have just already saved the datacube, and you
only now want to update this metadata.
HISTORY:
initial version imported GPItv (with definition of contrast corrected) - JM
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window; else = Window number to display in. |
| update_prev_fits_header | int | [0,1] | 0 | Update FITS metadata in the most recently saved datacube? |
| SaveProfile | string | None | Save radial profile to filename as FITS (blank for no save, dir name for default naming, AUTO for auto full path) | |
| SavePNG | string | None | Save plot to filename as PNG (blank for no save, dir name for default naming, AUTO for auto full path) | |
| contrsigma | float | [0.,20.] | Contrast sigma limit | |
| slice | int | [-1,50] | -1 | Slice to plot. -1 for all |
| DarkHoleOnly | int | [0,1] | 1 | 0: Plot profile in dark hole only; 1: Plot outer profile as well. |
| contr_yunit | int | [0,2] | 0 | 0: Standard deviation; 1: Median; 2: Mean. |
| contr_xunit | int | [0,1] | 0 | 0: Arcsec; 1: lambda/D. |
| yscale | int | [0,1] | 0 | 0: Auto y axis scaling; 1: Manual scaling. |
| contr_yaxis_type | int | [0,1] | 1 | 0: Linear; 1: Log |
| contr_yaxis_min | float | [0.,1.] | 0.00000001 | Y axis minimum |
| contr_yaxis_max | float | [0.,1.] | Y axis maximum |
IDL Filename: gpi_measure_contrast.pro
Plot the satellite spot locations vs. the expected location from wavelength scaling¶
Measured vs. wavelength scaled sat spot locations
Category: SpectralScience Order: 2.7
Inputs: Spectral mode datacube
Outputs: Plot of results.
Notes:
This is a quality check routine that verifies the expected wavelength
scaling of the datacube, based on how the satellite spots locations
vary with wavelength.
KEYWORDS:
HISTORY:
written 12/11/2012 - ds
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Display | int | [-1,100] | 1 | -1 = No display; 0 = New (unused) window; else = Window number to display in. |
| SaveData | string | None | Save data to filename (blank for no save) | |
| SavePNG | string | None | Save plot to filename as PNG(blank for no save) |
IDL Filename: gpi_meas_satspot_dev.pro
KLIP algorithm Spectral Differential Imaging¶
Reduce speckle noise using the KLIP algorithm across the spectral axis of a datacube.
Category: SpectralScience Order: 2.8
Inputs: Spectral datacube
Outputs: Spectral datacube after SDI PSF subtraction Output Suffix: Could not be determined automatically
Notes:
This algorithm reduces PSF speckles in a datacube using the
KLIP algorithm and Spectral Differential Imaging.
ALGORITHM:
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
HISTORY:
Written 2013. Tyler Barker
2013-07-18 MP: Renamed for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| refslice | int | [0,100] | 0 | Slice of the reference PSF |
| annuli | int | [0,100] | 5 | Number of annuli used |
| movement | float | [0.0,5.0] | 2.0 | Minimum pixel movement for reference set |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| arcsec | float | [0.0,1.0] | .4 | Radius of interest if using 1 annulus |
| signal | int | [0,1] | 0 | 1: calculate signal to noise ration, 0: don’t calculate |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_algorithm_spectral_differential_imaging.pro
Update World Coordinates¶
Add wcs info, assuming target star is precisely centered.
Category: ALL Order: 2.9
Inputs: Any datacube
Outputs: Datacube with updated WCS keywords in header Output Suffix: ‘-addwcs’ ; output suffix
Notes:
Creates a WCS-compliant header based on the target star's RA and DEC.
Currently assumes the target star is precisely centered.
KEYWORDS:
CalibrationFile= Name of astrometric binaries calibration file
GEM/GPI KEYWORDS:CRPA,RA,DEC
DRP KEYWORDS: CDELT1,CDELT2,CRPIX1,CRPIX2,CRVAL1,CRVAL2,CTYPE1,CTYPE2,HISTORY,PC1_1,PC2_2,PSFCENTX,PSFCENTY
HISTORY:
JM 2009-12
JM 2010-03-10: add handle of x0,y0 ref point with PSFcenter
2010-10-19 JM: split HISTORY keyword if necessary
2011-08-01 MP: Update for multi-extension FITS
2013-07-18 MP: Rename for consistency
2013-08-07 ds: idl2 compiler compatible
2013-12-30 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | string | None | AUTOMATIC | Name of astrometry offset calibratoin file |
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 0 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_update_world_coordinates.pro
Stores calibration in dataset¶
Stores the current calibration into the dataset structure.
Category: Calibration Order: 3.0
Inputs: data-cube
Outputs:
Notes:
To be called before an accumulate image.
It is used for high resolution microlens PSF determination
common needed:
KEYWORDS:
HISTORY:
Originally by Jean-Baptiste Ruffio 2013-08
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: store_calib_in_dataset.pro
Create Lenslet Flat Field¶
Create a 2D flat field for wavelength-independent lenslet throughput variations.
Category: Calibration Order: 3.1992
Inputs: Flat lamp data
Outputs: 2D lenslet flat field file Output Suffix: ‘lensletflat’
Notes:
Creates a simple derived flat field for non-uniform transmission in the
lenslets.
WARNING: experimental code, probably not yet ready for prime time.
HISTORY:
2014-01-02 MP: Created
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_lenslet_flat_field.pro
Smooth a 3D Cube¶
Smooth a cube by convolution with a Gaussian kernel, repeated for each slice of the cube.
Category: PolarimetricScience,SpectralScience,Calibration Order: 3.5
Inputs: polarimetry data-cube with flat lamp
Outputs: smoothed datacube Output Suffix: ‘clean’
Notes:
Convolves images with a gaussian kernel
Note: This primitive will work properly either before or after Accumulate
Images. If after, it will smooth all accumulated images.
GEM/GPI KEYWORDS:
DRP KEYWORDS:NAXES,NAXISi,FILETYPE,ISCALIB
HISTORY:
2014-01-09 MMB created
2014-04-28 MP: Minor documentation updates
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Smooth_FWHM | int | [0,100] | 3 | FWHM of gaussian kernel |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_smooth_cube.pro
Divide by Polarized Flat Field¶
Divides a 2-slice polarimetry file by a flat field.
Category: PolarimetricScience, Calibration Order: 3.5
Inputs: data-cube
Outputs: datacube with slices flat-fielded
Notes:
** Needs additional work, will not produce high qualty results yet **
GEM/GPI KEYWORDS:
DRP KEYWORDS: HISTORY
HISTORY:
2009-07-22: MDP created
2009-09-17 JM: added DRF parameters
2009-10-09 JM added gpitv display
2010-10-19 JM: split HISTORY keyword if necessary
2011-07-30 MP: Updated for multi-extension FITS
2013-07-12 MP: Rename for consistency
2013-12-16 MP: CalibrationFile argument syntax update.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| CalibrationFile | String | None | AUTOMATIC | Filename of the desired wavelength calibration file to be read |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_divide_by_polarized_flat_field.pro
Subtract Mean Stellar Polarization from podc¶
This description of the processing or calculation will show ; up in the Recipe Editor GUI. This is an example template for creating new ; primitives. It multiples any input cube by a constant value.
Category: PolarimetricScience Order: 3.85
Inputs: Coronagraphic mode Stokes Datacube
Outputs: That datacube with an estimated stellar polarization subtracted off. Output Suffix: ‘podc_sub’ ; set this to the desired output filename suffix
Notes:
Subtract an estimate of the stellar polarization, measured from
the mean polarization inside the occulting spot radius.
This primitive is simple, but has not been extensively tested.
Under what circumstances, if any, it is useful on GPI data in practice
is still TBD.
GEM/GPI KEYWORDS:
DRP KEYWORDS: HISTORY
HISTORY:
2014-03-23 MP: Started
2014-05-14 MMB: Rewrote for podc
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | String | Auto|Manual | Auto | Choose where to meausre the inst_pol? |
| InnerRadius | float | [-1,140] | -1 | The inner radius where you start to measure the instrumental polarization. -1 = the size of the FPM |
| OuterRadius | float | [0,140] | 20 | The radius out to which you measure the instrumental polarization. 0 = inside the occulter |
| WriteToFile | int | [0,1] | 0 | 1: Write the difference to a file, 0: Dont |
| Filename | string | None | Stellar_Polarization.txt | The filename where you write out the stellar polarization |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_subtract_mean_stellar_polarization_podc.pro
Rotate North Up¶
Rotate datacubes so that north is up and east is left.
Category: SpectralScience,PolarimetricScience Order: 3.9
Inputs: Datacube(s) in either spectral or polarimetric mode
Outputs: Rotated datacube(s) with north up and east left. Output Suffix: ‘-northup’
Notes:
Rotate so that North is Up, and east is to the left.
If necessary this will flip handedness as well as rotate
to get the right parity in the output image.
Note that this primitive can go *either* before or after
Accumulate Images. As a Level 1 primitive, it will rotate
one cube at a time; as a Level 2 primitive it will rotate
the whole stack of accumulated images all at once (though
it rotates each one by its own particular rotation angle).
KEYWORDS:
GEM/GPI KEYWORDS:RA,DEC,PAR_ANG
DRP KEYWORDS: CDELT1,CDELT2,CRPIX1,CRPIX2,CRVAL1,CRVAL2,NAXIS1,NAXIS2,PC1_1,PC1_2,PC2_1,PC2_2
HISTORY:
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Rot_Method | string | CUBIC|FFT | CUBIC | Method to compute the rotation |
| Center_Method | string | HEADERS|MANUAL | HEADERS | Determine the center of rotation from FITS header keywords, manual entry |
| centerx | int | [0,281] | 140 | Center X Pixel if Center_Method=Manual |
| centery | int | [0,281] | 140 | Center Y Pixel if Center_Method=Manual |
| pivot | int | [0,1] | 0 | Pivot about the center of the image? 0 = No |
| Save | int | [0,1] | 0 | None |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_rotate_north_up.pro
Rotate Field of View Square¶
Rotate datacubes so that the field of view is squarely aligned with the image axes.
Category: SpectralScience,PolarimetricScience Order: 3.9
Inputs: datacube
Outputs: Rotated datacube Output Suffix: ‘_rot’
Notes:
Rotate by the lenslet/field relative angle, so that the GPI IFS
field of view is roughly square with the pixel coordinate axes.
KEYWORDS:
GEM/GPI KEYWORDS:RA,DEC,PAR_ANG
DRP KEYWORDS: CDELT1,CDELT2,CRPIX1,CRPIX2,CRVAL1,CRVAL2,NAXIS1,NAXIS2,PC1_1,PC1_2,PC2_1,PC2_2
HISTORY:
2012-04-10 MDP: Created, based on rotate_north_up.pro
2013-11-07 ds - updated to use gpi_update_wcs_basic
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | enum | CUBIC|FFT | CUBIC | None |
| crop | int | [0,1] | 0 | Set to 1 to crop out non-illuminated pixels |
| Show | int | [0,1] | 0 | None |
| Save | int | [0,1] | 0 | None |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_rotate_field_of_view_square.pro
Accumulate Images¶
Stores images for combination by a subsequent primitive. Can buffer on disk for datasets too large to hold just in RAM.
Category: ALL Order: 4.0
Inputs: data-cube
Outputs: Output Suffix: save_currdata( DataSet, Modules[thisModuleIndex].OutputDir, suffix, display
Notes:
Stores images for later combination.
common needed:
KEYWORDS:
HISTORY:
2009-07-22 MDP: started
2009-09-17 JM: added DRF parameters
2013-08, 2013-10 MDP: Minor code formatting cleanup
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | string | OnDisk|InMemory | InMemory | OnDisk|InMemory |
IDL Filename: gpi_accumulate_images.pro
Create 2D Low Frequency Flat¶
Create Low Frequency Flat 2D from polarimetry flats
Category: HIDDEN Order: 4.01
Inputs: Polarimetry flats
Outputs: Low frequency flat Output Suffix: ‘-LFFlat’
Notes:
/!\ PROBABLY OUT OF DATE /!\ /!\ PROBABLY OUT OF DATE /!\
That's why it is hidden. It was not even ready for release.
See new LF flat determination with the work in progress by JB and Patrick on high resolution microlens PSF.
/!\ PROBABLY OUT OF DATE /!\ /!\ PROBABLY OUT OF DATE /!\
This primitive use a combined image of polarimetry flat-fields to build a low frequency flat for the detector array.
The idea applied here is to integrate the flux of each single spot using aperture photometry. For every spot, the neighbors are masked and the function aper computes the total flux and the sky/background correction. The neighbors are masked in order to have enough pixels to compute the sky. Then, we combine the flux of every couple of spots.
Then, we divide all the values with their median.
Artifacts are removed by computing a local std deviation and median. Every pixels further than n-sigma were replaced by the value of the smoothed flat.
To finish, we interpolates the resulting values over the 2048x2048 detector array using triangulate/trigrid function (linear interpolation using nearest neighbors).
There are still borders problem. The flat is not valid near the edges.
HISTORY:
Originally by Jean-Baptiste Ruffio 2013-06
2013-07-17 MP: Renamed for consistency
2013-12-03 MP: Add check for GCALLAMP=QH on input images
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_2d_low_frequency_flat.pro
Creates a thermal/sky background datacube¶
Create Sky/Thermal background cubes
Category: Calibration Order: 4.01
Inputs: A 2D sky image (should be a combination of several frames)
Outputs: A master sky frame, saved as a calibration file Output Suffix: ‘-bkgnd_cube’
Notes:
Create a thermal/sky background cube (3D) rather than using 2D detector frames as is done using the Combine 2D Thermal/Sky Backgrounds primitive. This allows a smoothing of the sky frame that will decrease the photon noise.
HISTORY:
2013-12-23 PI: Created Primitive
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| smooth_box_size | int | [0,100] | 3 | Size of box to smooth by (0: No smooth) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_sky_bkgd_cube.pro
Create microphonics noise model¶
Create a microphonics noise model in Fourier space.
Category: Calibration Order: 4.01
Inputs: several dark frames with strong microphonics
Outputs: microphonics model in Fourier space, saved as a calibration file Output Suffix: ‘-microModel’
Notes:
Create a microphonics noise model in Fourier space. The model consits only on the absolute value of the Fourier coefficients.
This has to be applied after the accumulate images primitive.
For each frame, it computes the absolute value of the Fourier coefficients around the 3 microphonics identified peaks.
Then it combines the results of all frames based on the method defined by Combining method.
Either adding all the models or taking the median.
Then it normalizes the model.
If Gauss_Interp = 1: Each of the 3 peaks is fitted by a 2d gaussian. Better not using it. It doesn't give better results.
HISTORY:
Originally by Jean-Baptiste Ruffio 2013-05
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Gauss_Interp | int | [0,1] | 0 | 1: Interpolate each peak by a 2d gaussian, 0: don’t interpolate |
| Combining_Method | string | ADD|MEDIAN | ADD | Method to combine the Fourier transforms of the microphonics (ADD|MEDIAN) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_microphonics_noise_model.pro
Combine 2D dark images¶
Combine 2D dark images into a master file via mean or median.
Category: Calibration Order: 4.01
Inputs: several dark frames
Outputs: master dark frame, saved as a calibration file Output Suffix: ‘-dark’
Notes:
Several dark frames are combined to produce a master dark file, which
is saved to the calibration database. This combination can be done using
either a mean or median algorithm, or a mean with outlier
rejection (sigma clipping)
Also, based on the variance between the various dark frames, the
read noise is estimated, and a list of hot pixels is derived.
The read noise and number of significantly hot pixels are written
as keywords to the FITS header for use in trending analyses.
HISTORY:
Jerome Maire 2008-10
2009-09-17 JM: added DRF parameters
2009-10-22 MDP: Created from mediancombine_darks, converted to use
accumulator.
2010-01-25 MDP: Added support for multiple methods, MEAN method.
2010-03-08 JM: ISCALIB flag for Calib DB
2011-07-30 MP: Updated for multi-extension FITS
2013-07-12 MP: Rename for consistency
2013-12-15 MP: Implemented SIGMACLIP, doc header updates.
2014-11-04 MP: Avoid trying to run parallelized sigmaclip if in IDL runtime.
2015-02-05 KBF: Fix readnoise estimation
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | string | MEAN|MEDIAN|SIGMACLIP | SIGMACLIP | How to combine images: median, mean, or mean with outlier rejection?[MEAN|MEDIAN|SIGMACLIP] |
| Sigma_cut | float | [1,100] | 3 | If Method=SIGMACLIP, then data points more than this many standard deviations away from the median value of a given pixel will be discarded. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_2d_dark_images.pro
Creates a thermal/sky background datacube¶
Create Sky/Thermal background cubes
Category: Calibration Order: 4.01
Inputs: A 2D sky image (should be a combination of several frames)
Outputs: A master sky frame, saved as a calibration file Output Suffix: ‘-bkgnd_cube’
Notes:
Create a thermal/sky background cube (3D) rather than using 2D detector frames as is done using the Combine 2D Thermal/Sky Backgrounds primitive. This allows a smoothing of the sky frame that will decrease the photon noise.
HISTORY:
2013-12-23 PI: Created Primitive
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| smooth_box_size | int | [0,100] | 3 | Size of box to smooth by (0: No smooth) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_sky_bkgd_cube.pro
Find Hot Bad Pixels from Darks¶
Find hot pixels from a stack of dark images (best with deep integration darks)
Category: Calibration Order: 4.01
Inputs: Multiple dark images
Outputs: Map of hot bad pixels Output Suffix: ‘-hotpix’
Notes:
This is a variant of combinedarkframes that (instead of combining darks)
analyzes them to find hot pixels and then writes out a
mask showing where the various hot pixels are.
The current algorithm determines pixels that are hot according to the
criteria:
(a) dark count rate must be > 1 e-/second for that pixel
(b) that must be measured with >5 sigma confidence above the estimated
read noise of the frames.
The first criterion can be adjusted using the hot_bad_thresh argument.
HISTORY:
2009-07-20 JM: created
2009-09-17 JM: added DRF parameters
2012-01-31 Switched sxaddpar to backbone->set_keyword Dmitry Savransky
2012-11-15 MP: Algorithm entirely replaced with one based on combinedarkframes.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| hot_bad_thresh | float | [0,100.] | 1.0 | Threshhold to consider a hot pixel bad, in electrons/second. |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_find_hot_bad_pixels_from_darks.pro
Find Cold Bad Pixels from Flats¶
Find cold pixels from a stack of flat images using all different filters
Category: Calibration Order: 4.01
Inputs: Flat field images, preferably in multiple filters
Outputs: Map of cold bad pixels Output Suffix: ‘-coldpix’
Notes:
This primitive finds cold (nonresponsive) pixels from the combination
of flat fields at multiple wavelengths. This trickier than one might think,
because it's not possible to illuminate the detector with an actual flat
field. The best you can do is a flat through the lenslet array but that's
still not very flat overall, with 37,000 spectra all over the place...
Instead, we can use a clever hack: let's add up flat fields at various
different wavelengths, and in the end we should get something that's actually
at least got some light into all the pixels. But there's a huge ripple pattern
everwhere.
We then rely on the symmetry of the lenslet array to compare a given pixel to
one that should have pretty similar flux levels, and we use that to find the
cold pixels.
HISTORY:
2013-03-08 MP: Implemented in pipeline based on algorithm from Christian
2013-12-03 MP: Add check for GCALLAMP=QH on input images
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_find_cold_bad_pixels_from_flats.pro
Create microphonics noise model¶
Create a microphonics noise model in Fourier space.
Category: Calibration Order: 4.01
Inputs: several dark frames with strong microphonics
Outputs: microphonics model in Fourier space, saved as a calibration file Output Suffix: ‘-microModel’
Notes:
Create a microphonics noise model in Fourier space. The model consits only on the absolute value of the Fourier coefficients.
This has to be applied after the accumulate images primitive.
For each frame, it computes the absolute value of the Fourier coefficients around the 3 microphonics identified peaks.
Then it combines the results of all frames based on the method defined by Combining method.
Either adding all the models or taking the median.
Then it normalizes the model.
If Gauss_Interp = 1: Each of the 3 peaks is fitted by a 2d gaussian. Better not using it. It doesn't give better results.
HISTORY:
Originally by Jean-Baptiste Ruffio 2013-05
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Gauss_Interp | int | [0,1] | 0 | 1: Interpolate each peak by a 2d gaussian, 0: don’t interpolate |
| Combining_Method | string | ADD|MEDIAN | ADD | Method to combine the Fourier transforms of the microphonics (ADD|MEDIAN) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_create_microphonics_noise_model.pro
Generate Combined Bad Pixel Map¶
This routine combines various sub-types of bad pixel mask (hot, cold, anomalous nonlinear pixels) to generate a master bad pixel list.
Category: Calibration Order: 4.02
Inputs: bad pixel maps
Outputs: Combined bad pixel map Output Suffix: ‘-badpix’
Notes:
This routine is used to combine the 3 individual types of bad pixel maps::
Hot bad pixels
Cold bad pixels
Nonlinear (too nonlinear to be usable) pixels
into one master bad pixel map.
This is an unusual recipe, in that its input file data is not actually
used in any way. All it does is use the first file to identify the date for
which the bad pixel maps are generated.
For this routine to run, there must be at least Hot and Cold bad pixel maps
already present in the calibration DB. The nonlinear pixels map is optional.
HISTORY:
Jerome Maire 2009-08-10
2009-09-17 JM: added DRF parameters
2012-01-31 Switched sxaddpar to backbone->set_keyword Dmitry Savransky
2012-11-19 MP: Complete algorithm overhaul.
2013-04-29 MP: Better error checking; nonlinearbadpix is optional.
2013-07-12 MP: rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_generate_combined_bad_pixel_map.pro
Clean Polarization Pairs via Double Difference¶
Category: PolarimetricScience,Calibration Order: 4.05
Inputs: Multiple polarization pair datacubes
Outputs: Multiple polarization pair datacubes, hopefully with reduced Output Suffix: Could not be determined automatically
Notes:
Given a sequence of polarization pair cubes, use a modified double
differencing approach to mitigate systematics between the e- and o- ray
channels of the cubes.
This must be used after Accumulate Images. Unlike most such primitives, it
acts on the entire stack of cubes at once without combining them yet.
This must be used prior to rotating the cubes if it is to have any hope at
all of working well.
**Caution** Experimental/Under Development code - algorithms may still be in
flux.
systematics
HISTORY:
2013-03-20 Started by Marshall, forked from gpi_combine_polarizations_dd.pro
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| fix_badpix | int | [0,1] | 1 | Also locate statistical outlier bad pixels and repair via interpolation? |
| Save_diffbias | int | [0,1] | 0 | Save the difference image systematic bias estimate subtracted from each pair? |
| gpitv_diffbias | int | [0,500] | 10 | Display empirical systematic bias in difference frames in a GPITV session 1-500, or 0 for no display |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| debug | int | [0,1] | 0 | Stop at breakpoints for debug/test |
IDL Filename: gpi_clean_polarization_pairs_dd.pro
Basic ADI¶
Implements the basic ADI algorithm described by Marois et al. (2006).
Category: SpectralScience Order: 4.1
Inputs:
Outputs: Output Suffix: Could not be determined automatically
Notes:
ADI algorithm based on original Marois et al (2006) paper.
KEYWORDS:
GEM/GPI KEYWORDS:FILTER,PAR_ANG,TELDIAM
DRP KEYWORDS: PSFCENTX,PSFCENTY
HISTORY:
Adapted for GPI - Jerome Maire 2008-08
multiwavelength - JM
2009-09-17 JM: added DRF parameters
2010-04-26 JM: verify how many spectral channels to process and adapt ADI for that,
so we can use ADI on collapsed datacubes or SDI
outputs
2013-08-07 ds: idl2 compiler compatible
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| numimmed | int | [1,100] | 3 | number of images for the calculation of the PSF reference |
| nfwhm | float | [0,20] | 1.5 | number of FWHM to calculate the minimal distance for reference calculation |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_basic_adi.pro
Primitive to interface TLOCI code and dependecies for PSF subtraction with the GPI pipeline.¶
Implements the TLOCI algorithm (Marois et al. 2014)
Category: SpectralScience Order: 4.11
Inputs: data-cubes spdc
Outputs: Not specified Output Suffix: ‘-tloci’
Notes:
KEYWORDS:
GEM/GPI KEYWORDS:
DRP KEYWORDS:
HISTORY:
ZHD: Intial creation.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| np | float | [0,20] | 1 | Number of processors to use. |
| badpix | int | [0,1] | 0 | Clean bad spaxels of each datacube? |
| unshrpmsk | int | [0,1] | 0 | Use unsharp mask? |
| register | int | [0,1] | 0 | Image registration? |
| p_spec | string | [STRING] | None | Planet spectrum filename. |
| s_spec | string | [STRING] | None | Stellar spectrum filename. |
| coeff | int | [0,1] | 0 | Positive Coeffecients? |
| Lambda | float | [0,1] | 0.5 | Lambda (Spectrum weighting) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_tloci.pro
ADI with LOCI¶
Implements the LOCI ADI algorithm (Lafreniere et al. 2007)
Category: SpectralScience Order: 4.11
Inputs: data-cube
Outputs: Not specified Output Suffix: Could not be determined automatically
Notes:
ADI algorithm based on Lafreniere et al. 2007.
Code currently only offers the use of positive and negative coefficients.
KEYWORDS:
GEM/GPI KEYWORDS:COADDS,CRFOLLOW,DEC,EXPTIME,HA,PAR_ANG
DRP KEYWORDS: HISTORY,PSFCENTX,PSFCENTY
HISTORY:
Jerome Maire :- multiwavelength 2008-08
JM: adapted for GPI-pip
2009-09-17 JM: added DRF parameters
2010-04-26 JM: verify how many spectral channels to process and adapt LOCI for that,
so we can use LOCI on collapsed datacubes or SDI outputs
2013-07-17 MP: Rename for consistency
2013-08-07 ds: idl2 compiler compatible, added start_primitive
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| nfwhm | float | [0,20] | 1.5 | number of FWHM to calculate the minimal distance for reference calculation |
| coeff_type | int | [0,1] | 0 | 0: positive and negative 1: positive only Coefficients in LOCI algorithm |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_adi_with_loci.pro
Populate Flexure Shifts vs Elevation Table¶
Derive shifts vs elevation lookup table.
Category: Calibration Order: 4.2
Inputs: Not specified
Outputs: Not specified Output Suffix: ‘-shifts’
Notes:
This function produces the table of spectral position shifts vs
elevation angle used to compensate for flexure within the GPI IFS.
It takes as input a series of reduced wavelength calibration
files. It compares them against a reference wavelength calibration file
obtained from the calibration database (which must have been taken in
horizontal orientation). The shift in X and Y positions are calculated
for each lenslet, and then the mean over the entire field is taken.
The median X shift and Y shift for each elevation is saved into
a table in the calibration database.
Optionally the user can request diagnostic plots displayed on screen or
saved to disk as postscript files.
HISTORY:
Jerome Maire 2013-02
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| Display | int | [-1,100] | 1 | -1 = No display; 0 = New (unused) window else = Window number to display diagnostic plot in. |
| saveplots | string | [yes|no] | no | Save diagnostic plots to PS files after running? [yes|no] |
IDL Filename: gpi_populate_flexure_shifts_vs_elevation_table.pro
Advanced KLIP ADI for Pol Mode¶
Reduce speckle noise using the KLIP algorithm with ADI data
Category: PolarimetricScience Order: 4.2
Inputs: Multiple spectral datacubes
Outputs: A reduced datacube with reduced PSF speckle halo Output Suffix: suffix+’-klip’
Notes:
This algorithm reduces PSF speckles in a datacube using the
KLIP algorithm and Angular Differential Imaging in Pol Mode
ALGORITHM:
Star location must have been previously measured using satellite spots.
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
HISTORY:
2013-10-21 - ds
2014-03-23 - MMB: Started adjusting for pol mode
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| annuli | int | [0,100] | 0 | Number of annuli to use |
| MinRotation | float | [0.0,360.0] | 1 | Minimum rotation between images (degrees) |
| CollapsePol | int | [0,1] | 0 | Collapse the pol cube and perform KLIP on the total intensity? |
| Mask | int | [0,1] | 0 | Do you want to mask out any area? If so you must provide a list of mask files |
| MaskList | string | None | MaskList.txt | The name of a file containing a list of mask files corresponding to each |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_adi_pol_adv.pro
KLIP ADI for Pol Mode¶
Reduce speckle noise using the KLIP algorithm with ADI data
Category: PolarimetricScience Order: 4.2
Inputs: Multiple spectral datacubes
Outputs: A reduced datacube with reduced PSF speckle halo Output Suffix: suffix+’-klip’
Notes:
This algorithm reduces PSF speckles in a datacube using the
KLIP algorithm and Angular Differential Imaging in Pol Mode
ALGORITHM:
Star location must have been previously measured using satellite spots.
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
HISTORY:
2013-10-21 - ds
2014-03-23 - MMB: Started adjusting for pol mode
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| annuli | int | [0,100] | 0 | Number of annuli to use |
| MinRotation | float | [0.0,360.0] | 1 | Minimum rotation between images (degrees) |
| CollapsePol | int | [0,1] | 0 | Collapse the pol cube and perform KLIP on the total intensity? |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_algorithm_angular_differential_imaging_pol.pro
Populate Flexure Shifts vs Elevation Table¶
Derive shifts vs elevation lookup table.
Category: Calibration Order: 4.2
Inputs: Not specified
Outputs: Not specified Output Suffix: ‘-shifts’
Notes:
This function produces the table of spectral position shifts vs
elevation angle used to compensate for flexure within the GPI IFS.
It takes as input a series of reduced wavelength calibration
files. It compares them against a reference wavelength calibration file
obtained from the calibration database (which must have been taken in
horizontal orientation). The shift in X and Y positions are calculated
for each lenslet, and then the mean over the entire field is taken.
The median X shift and Y shift for each elevation is saved into
a table in the calibration database.
Optionally the user can request diagnostic plots displayed on screen or
saved to disk as postscript files.
HISTORY:
Jerome Maire 2013-02
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| Display | int | [-1,100] | 1 | -1 = No display; 0 = New (unused) window else = Window number to display diagnostic plot in. |
| saveplots | string | [yes|no] | no | Save diagnostic plots to PS files after running? [yes|no] |
IDL Filename: gpi_populate_flexure_shifts_vs_elevation_table.pro
KLIP algorithm Angular Differential Imaging With Center Forced¶
Reduce speckle noise using the KLIP algorithm with ADI data with center forced to image center
Category: SpectralScience Order: 4.2
Inputs: Multiple spectral datacubes
Outputs: A reduced datacube with reduced PSF speckle halo Output Suffix: suffix+’-klip’
Notes:
This algorithm reduces noise in a datacube using the
KLIP algorithm and Angular Differential Imaging.
This is the same as the 'KLIP algorithm Angular Differential Imaging'
primitive, except the location of the star can be entered manually.
This can be useful for data with low S/N where the satellite
spots are not found automatically.
Used during first light run, may not be that generally applicable
now that automatic satellite spot finding is much more robust.
ALGORITHM:
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
HISTORY:
2013-11-16- ds Developed on the fly during first light run
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| annuli | int | [0,100] | 0 | Number of annuli to use |
| centerx | int | [0,281] | 140 | Center X Pixel |
| centery | int | [0,281] | 140 | Center Y Pixel |
| MinRotation | float | [0.0,360.0] | 1 | Minimum rotation between images (degrees) |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_algorithm_angular_differential_imaging_forcecent.pro
KLIP ADI for Pol Mode¶
Reduce speckle noise using the KLIP algorithm with ADI data
Category: PolarimetricScience Order: 4.2
Inputs: Multiple spectral datacubes
Outputs: A reduced datacube with reduced PSF speckle halo Output Suffix: suffix+’-klip’
Notes:
This algorithm reduces PSF speckles in a datacube using the
KLIP algorithm and Angular Differential Imaging in Pol Mode
ALGORITHM:
Star location must have been previously measured using satellite spots.
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
HISTORY:
2013-10-21 - ds
2014-03-23 - MMB: Started adjusting for pol mode
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| annuli | int | [0,100] | 0 | Number of annuli to use |
| MinRotation | float | [0.0,360.0] | 1 | Minimum rotation between images (degrees) |
| CollapsePol | int | [0,1] | 0 | Collapse the pol cube and perform KLIP on the total intensity? |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_algorithm_angular_differential_imaging_pol.pro
Combine Wavelength Calibrations¶
Performs simple median combination of wavelength calibrations from flat and/or arc lamps
Category: Calibration Order: 4.2
Inputs: Multiple 3D wavcal cubes
Outputs: One merged 3D wavecal cube
Notes:
gpi_combine_wavcal_all is a simple median combination of wav. cal. files obtained with flat and arc images.
TO DO: exclude some mlens from the median in case of wavcal
This is **mostly deprecated**: in general it is recommended to combine
all 2D images for a given arc lamp and then derive one wavelength solution
from those, rather than deriving multiple wavecals and then combining them.
On the other hand if you want to merge wavecals from two different lamps then
you can indeed use this.
GEM/GPI KEYWORDS:DATE-OBS,FILTER,IFSFILT,TIME-OBS
DRP KEYWORDS: DATAFILE
HISTORY:
Jerome Maire 2009-08-10
2009-09-17 JM: added DRF parameters
2012-10-17 MP: Removed deprecated suffix= keyword
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_wavelength_calibrations.pro
KLIP algorithm ADI + SDI¶
Reduce speckle noise using the KLIP algorithm with ADI+SDI data
Category: SpectralScience Order: 4.2
Inputs: Multiple spectral datacubes, ADR corrected
Outputs: A reduced datacube with reduced PSF speckle halo Output Suffix: Could not be determined automatically
Notes:
This algorithm reduces PSF speckles in a datacube using the
KLIP algorithm and Angular Differential Imaging + Spectral
Differential Imaging. Based on Soummer et al., 2012.
Star location must have been previously measured using satellite spots.
Measure annuli out from the center of the cube and create a
reference set for each annuli of each slice. Apply KLIP to the
reference set and project the target slice onto the KL
transform vector. Subtract the projected image from the
original and repeat for all slices
*** Development code, not intended for public release ***
*** Does not work in compiled mode, but that's OK since not intended for
public release ****
HISTORY:
2014-05 JW modified code from Dmitry and Tyler Barker
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: save output on disk, 0: don’t save |
| numthreads | int | [1,10000] | 5 | Number of parallel processes to run KLIP calculation |
| annuli | int | [0,100] | 5 | Number of annuli to use |
| subsections | int | [1,10] | 4 | Number of equal area subsections to break up each annulus into |
| prop | float | [0.8,1.0] | .99999 | Proportion of eigenvalues used to truncate KL transform vectors |
| minsep | float | [0.0,250] | 3 | Minimum separation between slices (pixels) |
| minPA | float | [0.0,360] | 0.0 | Minimum parallactic rotation (in degrees) for constructing reference PSF (good for disk targets) |
| waveclip | int | [0,18] | 2 | Number of wavelength slices at the beginning and end of each cube to ignore |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_klip_adi_plus_sdi.pro
Simple SDI of post ADI residual¶
Apply SSDI to create a 2D subtracted image from a cube. Given the user’s specified wavelength ranges, extract the 3D datacube slices for each of those wavelength ranges. Collapse these down into 2D images by simply averaging the values at each lenslet (ignoring NANs). Rescale Image1, then compute diffImage = I1scaled - k* I2
Category: SpectralScience Order: 4.3
Inputs: Datacube of ADI residuals
Outputs: Datacube with simple 2-wavelength SDI subtraction
Notes:
This recipe rescales and subtracts 2 frames in different user-defined
bandwidths. This recipe is used for speckle suppression using the
Marois et al (2000) algorithm.
This routine does NOT update the data structures in memory. You **MUST**
set the keyword SAVE=1 or else the output is silently discarded.
ALGORITHM:
Given the user's specified wavelength ranges, extract the 3D datacube slices
for each of those wavelength ranges. Collapse these down into 2D images by
simply averaging the values at each lenslet (ignoring NANs). Rescale Image1
using fftscale so that the PSF scale matches that of Image2 (as computed
from the average wavelength for each image). Then compute
diffImage = I1scaled - k* I2
Then hopefully output the image somewhere if SAVE=1 is set.
HISTORY:
2007-11 Jerome Maire
2009-04-15 MDP: Documentation updated; slight code cleanup
2009-09-17 JM: added DRF parameters
2013-07-17 MP: Renamed for consistency
2013-08-07 ds: idl2 compiler compatible
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| L1Min | float | [0.9,2.5] | 1.55 | Wavelength range 1, minimum wavelength [in microns] |
| L1Max | float | [0.9,2.5] | 1.57 | Wavelength range 1, maximum wavelength [in microns] |
| L2Min | float | [0.9,2.5] | 1.60 | Wavelength range 2, minimum wavelength [in microns] |
| L2Max | float | [0.9,2.5] | 1.65 | Wavelength range 2, maximum wavelength [in microns] |
| k | float | [0,10] | 1.0 | Scaling factor of Intensity(wav_range2) with diffImage = I1scaled - k* I2 |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 5 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_simple_sdi_of_post_adi_residual.pro
Combine Polarization Sequence via Double Difference¶
Category: PolarimetricScience,Calibration Order: 4.4
Inputs: Multiple polarization pair datacubes
Outputs: a single Stokes datacube Output Suffix: Could not be determined automatically
Notes:
Combine a sequence of polarized images via the SVD method, after first
performing double differencing to remove systematics between the e- and
o-rays.
See James Graham's SVD algorithm document, or this algorithm may be hard to
follow. This is not your father's imaging polarimeter any more!
This routine assumes that it can read in a series of files on disk which were written by
the previous stage of processing.
GEM/GPI KEYWORDS:EXPTIME,ISS_PORT,PAR_ANG,WPANGLE
DRP KEYWORDS:CDELT3,CRPIX3,CRVAL3,CTYPE3,CUNIT3,DATAFILE,NAXISi,PC3_3
HISTORY:
2009-07-21: MDP Started
2009-09-17 JM: added DRF parameters
2013-01-30: updated with some new keywords
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| HWPoffset | float | [-360.,360.] | -29.14 | The internal offset of the HWP. If unknown set to 0 |
| IncludeSystemMueller | int | [0,1] | 1 | 1: Include, 0: Don’t |
| IncludeSkyRotation | int | [0,1] | 1 | 1: Include, 0: Don’t |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_polarizations_dd.pro
Combine Polarization Sequence¶
Category: PolarimetricScience,Calibration Order: 4.4
Inputs:
Outputs: Not specified Output Suffix: “-stokesdc”
Notes:
Combine a sequence of polarized images via the SVD method.
See James Graham's SVD algorithm document, or this algorithm may be hard to
follow. This is not your father's imaging polarimeter any more!
This routine assumes that it can read in a series of files on disk which were written by
the previous stage of processing.
GEM/GPI KEYWORDS:EXPTIME,ISS_PORT,PAR_ANG,WPANGLE
DRP KEYWORDS:CDELT3,CRPIX3,CRVAL3,CTYPE3,CUNIT3,DATAFILE,NAXISi,PC3_3
ALGORITHM:
HISTORY:
2009-07-21: MDP Started
2009-09-17 JM: added DRF parameters
2013-01-30: updated with some new keywords
2014-03 MP and MM-B: Polarization coordinates and angles verification and debug
2014-08-18 MPF: some formatting cleanup, added sanity check, fixed
which header was used for output file
2014-03 MP and MM-B: Polarization coordinates and angles verification and debug
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| HWPoffset | float | [-360.,360.] | -29.14 | The internal offset of the HWP. If unknown set to 0 |
| IncludeSystemMueller | int | [0,1] | 0 | 1: Include, 0: Don’t |
| IncludeSkyRotation | int | [0,1] | 1 | 1: Include, 0: Don’t |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_polarization_sequence.pro
Combine 3D Datacubes¶
Combine 3D datacubes via mean or median.
Category: ALL Order: 4.5
Inputs: 3d datacubes
Outputs: a single combined datacube Output Suffix: strlowcase(method)
Notes:
Multiple 3D cubes can be combined into one, using either a Mean or a Median.
TODO: more advanced combination methods. Improved sigma-clipped mean implementation
HISTORY:
Jerome Maire 2008-10
2009-09-17 JM: added DRF parameters
2009-10-22 MDP: Created from mediancombine_darks, converted to use
accumulator.
2010-01-25 MDP: Added support for multiple methods, MEAN method.
2011-07-30 MP: Updated for multi-extension FITS
2012-10-10 MP: Minor code cleanup
2013-07-29 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Method | enum | MEAN|MEDIAN|SIGMACLIP|MINIMUM | MEDIAN | How to combine images: median, mean, or mean with outlier rejection? |
| sig_clip | int | 0,10 | 3 | Clipping value to be used with SIGMACLIP in sigma (stddev) |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_combine_3d_datacubes.pro
Quality Check Wavelength Calibration¶
Performs a basic quality check on a wavecal based on the statistical distribution of measured inter-lenslet spacings.
Category: Calibration Order: 4.5
Inputs: 3D wavecal
Outputs: if wavecal fails quality check, then recipe is failed
Notes:
Quality check wavelength calibration:
Looks for unexpected statistical anomalies or offsets in the
wavelength calibration, as implemented in the utility
function `gpi_wavecal_sanity_check`
HISTORY:
2013-11-28 MP: Created.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| error_action | string | [Fail|Ask_user] | Ask_user | If the quality check fails, should the recipe immediately fail or should I alert the user and ask what they want to do? |
| Display | int | [-1,100] | -1 | -1 = No display; 0 = New (unused) window; else = Window number to display diagnostic plot. |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
IDL Filename: gpi_quality_check_wavelength_calibration.pro
Median Combine ADI datacubes¶
Median combine all residual datacubes after an ADI (or LOCI) speckle suppression.
Category: SpectralScience Order: 4.5
Inputs: Many datacubes with ADI subtraction residuals
Outputs: Median combined residual datacube Output Suffix: Could not be determined automatically
Notes:
Median all ADI datacubes
HISTORY:
Jerome Maire :- multiwavelength 2008-08
JM: adapted for GPI-pip
2009-09-17 JM: added DRF parameters
2010-10-19 JM: split HISTORY keyword if necessary
2013-07-16 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 10 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_median_combine_adi_datacubes.pro
Pad Wavelength Calibration Edges¶
pads the outer edges of the wavecal via extrapolation to cover lenslets whose spectra only partially fall on the detector field of view.
Category: Calibration Order: 4.6
Inputs: 3D wavecal
Outputs: 3D wavecal with extrapolated edges
Notes:
This primitive extrapolates from the lenslets near the edge of the
detector to provide estimated wavelength solutions for the 'fractional'
lenslets, whose spectra only fall partially on the detector.
Note that these fractional spectra are not generally going to be
useful for scientific analyses, but for some datacube processing
tasks such as destriping it can be helpful to have wavelength
solutions that cover these lenslets.
HISTORY:
2013-11-28 MP: Created.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 0 | 1: Save output to disk, 0: Don’t save |
| gpitvim_dispgrid | int | [0,500] | 15 | 1-500: choose gpitv session for displaying image output and wavcal grid overplotted, 0: no display |
IDL Filename: gpi_pad_wavelength_calibration_edges.pro
Flexure Quicklook for Spectra (Lsqr, microlens psf)¶
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square algorithm and microlenslet PSFs. Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Category: SpectralScience Order: 5.0
Inputs: 2D detector image, wavecal, microlens PSF reference.
Outputs: None Output Suffix: ‘quick_residual’ ; set this to the desired output filename suffix
Notes:
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square, matrix inversion algorithm and microlenslet PSFs.
Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Ideally run in parrallel enviroment.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| x_lens | float | [0,281] | 150 | Lenslet number in x so start search |
| y_lens | float | [0,281] | 150 | Lenslet number in y so start search |
| size | float | [0,30] | 5 | Size of region to iterate |
| resid | float | [0,1] | 1 | Save residual detector image? |
| micphn | float | [0,1] | 0 | Solve for microphonics? |
| iter | float | [0,2] | 1 | Run iterative solver of wavecal? |
| badpix | float | [0,1] | 0 | Weight by bad pixel map? |
| del_x_best | float | [-5,5] | 0 | Best initial guess for flexure perpendicular to dispersion shift (pixels) |
| del_theta_best | float | [-5,5] | 0 | Best initial guess for rotation angle shift (degrees) |
| del_lam_best | float | [-5,5] | 0 | Best initial guess for flexure parrallel to dispersion shift (pixels) |
| x_off | float | [-5,5] | 0 | Offset from wavecal in x pixels (used only if xcorrelariotn wasn’t run!) |
| y_off | float | [-5,5] | 0 | Offset from wavecal in y pixels |
IDL Filename: gpi_lsqr_mlens_quick_flx.pro
Flexure Quicklook for Pol (Lsqr, microlens psf)¶
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square algorithm and microlenslet PSFs. Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Category: PolarimetricScience Order: 5.0
Inputs: 2D detector image, wavecal, microlens PSF reference.
Outputs: None Output Suffix: ‘’ ; set this to the desired output filename suffix
Notes:
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square, matrix inversion algorithm and microlenslet PSFs.
Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Ideally run in parrallel enviroment.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| x_lens | float | [0,281] | 150 | Lenslet number in x so start search |
| y_lens | float | [0,281] | 150 | Lenslet number in y so start search |
| size | float | [0,30] | 4 | Size of region to iterate |
| resid | float | [0,1] | 1 | Save residual detector image? |
| micphn | float | [0,1] | 0 | Solve for microphonics? |
| iter | float | [0,1] | 1 | Run iterative solver of wavecal? |
| del_x | float | [-5,5] | 0 | Best initial guess for flexure in detector x (pixels) |
| del_y | float | [-5,5] | 0 | Best initial guess for flexure in detector y (pixels) |
| x_off | float | [-5,5] | 0 | Offset from wavecal in x pixels (used only if xcorrelariotn wasn’t run!) |
| y_off | float | [-5,5] | 0 | Offset from wavecal in y pixels |
IDL Filename: gpi_lsqr_mlens_pol_quick_flx.pro
Subtract Mean Stellar Polarization¶
This description of the processing or calculation will show ; up in the Recipe Editor GUI. This is an example template for creating new ; primitives. It multiples any input cube by a constant value.
Category: PolarimetricScience Order: 5.0
Inputs: Coronagraphic mode Stokes Datacube
Outputs: That datacube with an estimated stellar polarization subtracted off. Output Suffix: ‘sub’
Notes:
Subtract an estimate of the stellar polarization, measured from
the mean polarization inside the occulting spot radius.
This primitive is simple, but has not been extensively tested.
Under what circumstances, if any, it is useful on GPI data in practice
is still TBD.
HISTORY:
2014-03-23 MP: Started
2015-02-05 LWH: Added more parameters
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_subtract_mean_stellar_polarization.pro
Assemble Polarization Datacube (Lsqr, microlens psf)¶
This primitive will extract flux from a 2D detector image into a GPI polarization cube using a least-square algorithm and microlenslet PSFs. Optionally can produce a residual detector image, solve for microphonics, and iterate the polcal solution to find a minimum residual.
Category: PolarimetricScience Order: 5.0
Inputs: 2D detector image, polcal, microlens PSF reference.
Outputs: GPI datacube Output Suffix: ‘-podc’ ; set this to the desired output filename suffix
Notes:
This primitive will extract flux from a 2D detector image of Wollaston spots into a GPI polarization cube using a least-square, matrix inversion algorithm and microlenslet PSFs.
Optionally can produce a residual detector image, solve for microphonics, and iterate the polcal solution to find a minimum residual.
Ideally run in parrallel enviroment.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-02-17 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| stopidl | int | [0,1] | 0 | 1: stop IDL, 0: dont stop IDL |
| np | float | [0,100] | 2 | Number of processors to use in reduction (double check enviroment before running) |
| resid | int | [0,1] | 1 | Save residual detector image? |
| micphn | int | [0,1] | 0 | Solve for microphonics? |
| iter | int | [0,1] | 1 | Run iterative solver of polcal? |
| x_off | float | [-5,5] | 0 | Offset from wavecal in x pixels |
| y_off | float | [-5,5] | 0 | Offset from wavecal in y pixels |
IDL Filename: gpi_lsqr_mlens_extract_pol.pro
Assemble Spectral Datacube (Lsqr, microlens psf)¶
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square algorithm and microlenslet PSFs. Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Category: SpectralScience Order: 5.0
Inputs: 2D detector image, wavecal, microlens PSF reference.
Outputs: GPI datacube Output Suffix: ‘spdc’ ; set this to the desired output filename suffix
Notes:
This primitive will extract flux from a 2D detector image into a GPI spectral cube using a least-square, matrix inversion algorithm and microlenslet PSFs.
Optionally can produce a residual detector image, solve for microphonics, and iterate the wavecal solution to find a minimum residual.
Ideally run in parrallel enviroment.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| np | float | [0,100] | 4 | Number of processors to use in reduction (double check enviroment before running) |
| resid | int | [0,1] | 1 | Save residual detector image? |
| micphn | int | [0,1] | 0 | Solve for microphonics? |
| iter | int | [0,1] | 0 | Run iterative solver of wavecal? |
| badpix | float | [0,1] | 0 | Weight by bad pixel map? |
| del_x_best | float | [-5,5] | 0 | Best initial guess for flexure perpendicular to dispersion shift (pixels) |
| del_theta_best | float | [-5,5] | 0 | Best initial guess for rotation angle shift (degrees) |
| del_lam_best | float | [-5,5] | 0 | Best initial guess for flexure parrallel to dispersion shift (pixels) |
| x_off | float | [-5,5] | 0 | Offset from wavecal in x pixels |
| y_off | float | [-5,5] | 0 | Offset from wavecal in y pixels |
IDL Filename: gpi_lsqr_mlens_extract.pro
Flexure 2D x correlation with wavecal model (perpendicular)¶
This primitive uses the relevent microlense PSF and wave cal to generate a model detector image to cross correlate with a science image.
Category: SpectralScience Order: 5.0
Inputs: Science image, microlens PSF, wavecal
Outputs: Flexure offset in xy detector coordinates. Output Suffix: ‘’ ; set this to the desired output filename suffix
Notes:
This primitive uses the relevent microlense PSF and wave cal to generate a model detector image to cross correlate with a science image.
The resulting output can be used as a flexure offset prior to flux extraction.
The resulting output can be used as a flexure offset prior to flux extraction.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| range | float | [0,5] | 2 | Range of cross corrleation search in pixels. |
| resolution | float | [0,1] | 0.01 | Subpixel resolution of cross correlation convergence |
| psf_sep | float | [0,1] | 0.01 | PSF separation in pixels |
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| del_x_best | float | [-5,5] | 0 | Best initial guess for flexure perpendicular to dispersion shift (pixels) |
| badpix | float | [0,1] | 0 | Weight by bad pixel map? |
IDL Filename: gpi_img_xcorr_perp.pro
Flexure 2D x correlation with ulens and polcal models¶
This primitive uses the relevent microlense PSF and pol cal to generate a model detector image to cross correlate with a science image and find the flexure offset.
Category: PolarimetricScience Order: 5.0
Inputs: Science image, microlens PSF, wavecal
Outputs: Flexure offset in xy detector coordinates. Output Suffix: ‘’ ; set this to the desired output filename suffix
Notes:
This primitive uses the relevent microlense PSF and pol cal to generate a model detector image to cross correlate with a science image.
The resulting output can be used as a flexure offset prior to flux extraction.
The resulting output can be used as a flexure offset prior to flux extraction.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
2014-10-01 MMB: Changed name to distinguish from other version that JUST uses polcal file.
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| range | float | [0,5] | 2 | Range of cross corrleation search in pixels. |
| resolution | float | [0,1] | 0.01 | Subpixel resolution of cross correlation |
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| x_off | float | [-5,5] | 0 | initial guess for large offsets |
| y_off | float | [-5,5] | 0 | initial guess for large offsets |
| badpix | float | [0,1] | 0 | Weight by bad pixel map? |
IDL Filename: gpi_img_xcorr_pol.pro
Flexure 2D x correlation with wavecal model¶
This primitive uses the relevent microlense PSF and wave cal to generate a model detector image to cross correlate with a science image.
Category: SpectralScience Order: 5.0
Inputs: Science image, microlens PSF, wavecal
Outputs: Flexure offset in xy detector coordinates. Output Suffix: ‘’ ; set this to the desired output filename suffix
Notes:
This primitive uses the relevent microlense PSF and wave cal to generate a model detector image to cross correlate with a science image.
The resulting output can be used as a flexure offset prior to flux extraction.
The resulting output can be used as a flexure offset prior to flux extraction.
where in the order of the primitives should this go by default?
pick one of the following options for the primitive type:
HISTORY:
Began 2014-01-13 by Zachary Draper
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| range | float | [0,5] | 2 | Range of cross corrleation search in pixels. |
| resolution | float | [0,1] | 0.01 | Subpixel resolution of cross correlation |
| yoff_factor | float | [0,5] | 2 | Fractor of resolution to which the code converges in y_offset |
| psf_sep | float | [0,1] | 0.01 | PSF separation in pixels |
| stopidl | int | [0,1] | 1 | 1: stop IDL, 0: dont stop IDL |
| x_off | float | [-5,5] | 0 | initial guess for large offsets |
| y_off | float | [-5,5] | 0 | initial guess for large offsets |
| badpix | float | [0,1] | 0 | Weight by bad pixel map? |
IDL Filename: gpi_img_xcorr.pro
Insert Planet into datacube¶
This primitive inserts planets into reduced datacubes. It can be run multiple times to insert multiple planets.
Category: SpectralScience Order: 5.0
Inputs: A fully reduced datacube prior to any speckle manipulation. A planet’s distance, separation, position angle, mass, age, and formation scenario (hot/cold start).
Outputs: The datacube with an inserted planet. Output Suffix: ‘wplnt’ ; set this to the desired output filename suffix
Notes:
This primitive allows users to input artificial planets, based on the solar
metalicity, hybrid cloud model, hot/cold formation scenario models of Spiegel
and Burrows (2012) into reduced datacubes. The planet PSF is represented by an
average of the four satellite spots. The models span the ages of 1 to 1000
Myr, and masses of 1-15 Jupiter masses. If a user specifies parameters that do
not represent an exact model the nearest model in age, then mass, is used.
Currently, the intensity of the planets is determined assuming an instrument
throughput of 18.6%, combined with a 7.9 meter primary mirror with a 1m
secondary. The user also has the option to scale the image to represent a star
of a user-defined magnitude. This provides the ability to simulate multiple
observing scenarios.
The stellar and planet properties are written to the headers. Should the user
wish to not include the planet information, it can be bypassed by
de-activating the write_header_info keyword.
At the moment there is no way to determine only the star parameters and not
insert a planet. To do this, the user should just put the planet distance to a
large number and separation to a small number.
Note that the inserted separation and position angle will be SLIGHTLY different
from the user specified values - the proper values can be found in the header
HISTORY:
2013-07-30 PI: Created Primitive
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| Age | int | [1,1000] | 10 | Age of planet in Myr |
| Mass | int | [1,15] | 10 | Mass of planet in Jupiter masses |
| model_type | string | [hot,cold] | hot | Hot or Cold Start formation scenario |
| position_angle | float | [0,360] | 45.0 | Position angle of the planet in degrees East of North |
| Separation | float | [0,1800] | 500 | Separation in milli-arcseconds |
| Star_Mag | float | [-1,8] | -1 | Stellar Magnitude in H band, -1 estimates stellar magnitude from satellite spots |
| distance | float | [0,1000] | 10.0 | distance to system in parsecs |
| write_header_info | int | [0,1] | 1 | 1: Write planet info to headers 0: don’t write planet info to headers |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_insert_planet_into_cube.pro
Save Accumulated Stack¶
Save output to disk as a FITS file. Note that you can often do this from another module by setting the ‘save=1’ option; this is a redundant way to specify that.
Category: ALL Order: 10.0
Inputs: Any
Outputs: The input is written to disk as a FITS file Output Suffix: Could not be determined automatically
Notes:
Save the current accumulated stack of images to disk.
Note that this uses whatever the currently defined suffix is, though you can
also override that here. This is the one and only routine that should be
used to override a suffix.
TODO: change output filename too, optionally?
HISTORY:
2014-11-18 Created by MMB
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| suffix | string | None | None | choose the suffix |
IDL Filename: gpi_save_accumulated_stack.pro
Save Output¶
Save output to disk as a FITS file. Note that you can often do this from another module by setting the ‘save=1’ option; this is a redundant way to specify that.
Category: ALL Order: 10.0
Inputs: Any
Outputs: The input is written to disk as a FITS file Output Suffix: Could not be determined automatically
Notes:
Save the current file to disk. Note that you can often do this
from another primitive by setting the 'save=1' option; this is an
optional, redundant way to specify that.
Note that this uses whatever the currently defined suffix is, though you can
also override that here. This is the one and only routine that should be
used to override a suffix.
TODO: change output filename too, optionally?
HISTORY:
2009-07-21 Created by MDP.
2009-09-17 JM: added DRF parameters
2013-07-17 MP: Rename for consistency
Parameters:
| Name | Type | Range | Default | Description |
|---|---|---|---|---|
| suffix | string | None | None | choose the suffix |
| Save | int | [0,1] | 1 | 1: save output on disk, 0: don’t save |
| gpitv | int | [0,500] | 2 | 1-500: choose gpitv session for displaying output, 0: no display |
IDL Filename: gpi_save_output.pro