Stardust actions and decisions


RAW WILD 2 DATA

-- Figured out how to recover the overscan pixels.  
      - Recovering them, but not the "sync word" or "line number" columns
      - Gives 8 pixels at the start of the line and 12 at the end
      - makes image 1044x1024
      - Includes the "fat pixels" on the sides of the image

-- For windowed image 
      - Inserting the windows into a full-frame image 
        - pixels outside windows are set to zero
      - Include the overscan pixels set to zero
      - Include the pixel locations defining the windows 
          - As part of the image description in the PDS label
          - As keywords in the FITS header
      - Window positions reflect the added overscan regions

-- Need to flip the raw data so the image appears as it would to
   an in situ observer
   - To preserve the overscan positions, flip vertically about a
     horizontal axis
   - Recalculated all angles and pixel positions in the header
   - Double/triple checked the geometry



-- Shutter timing correction

   - As part of Stardust NExT calibrations, Ken determined that the
     shutter direction is set to forward when the NAVCAM is powered off.
     Got a list of all times when power is turned off and on.  Use it to
     help figure out the shutter directions

   - Parity still changes when a bias frame is taken (shutter doesn't
     trigger.  Have a list of bias frames (only a few).

   - Propose to only correct shutter exposure times where they are
     less than 1000 msec.
     -- Many occasions when missing images introduce uncertainty
        into shutter direction.  These uncertanties only affect longer
        exposures
     -- For consistency, don't fix any of the longer exposures
     -- The longer exposures are insensitive to the error 
     -- The shorter exposures seem to have no ambiguities about direction
        (pending the analysis by JYL)
     -- Document what was done and why
        (this will add several documents to all data sets)

   - AnneFrank data is also affected. 
     -- Team applied corrections according to what they though was happening
     -- Power resets make some of these corrections wrong
     -- Re-do the AnneFrank exposures to be consistent, following the
        same rationale as discussed for Wild 2 data.





CALIBRATED WILD 2 DATA

-- Overscan regions were removed in the image processing, and are not
   included in the IMG data files
   -- could grab them from the raw images, but I don't see any reason
      to do that 

-- "Fat pixels" were also removed 

-- Need to correct for the exposure time changes
   - not applying Ken's new calibrations (retain the Stardust team's
     calibrations, even though they may not be correct) 
   - That is what we got from the team
   - not clear the new calibrations are applicable for the Wild 2 timeframe
   - Add a note to the documentation that there is a new calibration
     coming, and the user can go apply it to the raw data if so
     desired
   - or that will be version 3 of the data


As yet unresolved issues:

-- Window information is not given in Calibrated image labels.  Should
   I grab window information from raw data and include it with the
   calibrated data?

-- I previously rotated the raw images to match the calibrated image
   orientation.  With the new flip that preserves the overscan region,
   the orientations are different.  Should I rotate the calibrated
   images to match the raw frame orientation 
   -- Discussion about orientation in a note in the labels.  I need to
      figure out how to edit it.

-- Is all this worth it?


GENERAL

--All actions/changes are being done via Perl scripts and IDL programs
  -- Always start with original IMG files
       - preserving them intact

  -- Makes it easier to re-run if changes are needed
  -- Documented what needs run and in what order
  -- When additional data are required (e.g. geometry) they are
     read in from data files, listed along with the image name
     - gives a record of what was used and allows for easy changes




Shutter correction

It looks like the Annefrank data have even frames that are shorter
than commanded, while in the Wild 2 data, the odd frames are shorter.
Not clear where the parity change cames from

Need to correct for the exposure time in calibrated data.



in wild2 dataset.cat file, in images 886-889 section: need to change
"Image 989" to "Image 889"






Additions/changes

Added file SHUTTER_CORRECTIONS.ASC to documents directory (+label)
Added file SHUTTER_PARITY_LIST.ASC to documents directory (+label)
Added PIXEL_SCALE_ANALYSIS.ASC to documents directory (+label)
Added FLAT_FIELD_ANALYSIS.ASC  to documents directory (+label)
   images PIXSCL.PNG and FIELD_DISTORT.PNG
Added NAVCAM_ON_OFF_TIMES.ASC to documents directory (+label)

Chaeck to make sure catalog files match between datasets


Recovered the overscan pixels in the raw images



all 3 DATASET.CAT files have been updated for the respective issues
involved.