Astrodrizzle CR rejection

zhao13zhao13 Member Posts: 8

So I was trying to make a drizzled image with 40 WFC-3(F814W) FLT files. For the step of CR rejection, I used 'driz_cr_snr = 3.5 3.0', drizzle_cr_scale=1.2, 0.7, driz_cr_grow = 1, driz_cr_ctegrow=0. However, when I checked the final output, there are still some "spikes" extending from both sides of some bright sources.

I checked each individual _blt and _crclean image, and found the CR masks work pretty well. Therefore, I wonder if this is affected by cosmic ray or some other effects (CTI trails, source saturation, etc.)

If it is indeed because of cosmic ray, can anyone give me any suggestions on the parameters I should have used? I attached the image I got for your reference.


Best regards,


  • Member, Institute_Staff, Moderator_DrizzlePac Posts: 28
    edited July 5
    These spikes are due to bleeding along columns from saturated stars. They should be flagged in the DQ array of the FLT frames.
    We encourage you to use FLC data products instead of FLT, since faint sources near the center of the detector will be heavily trailed due to charge transfer losses. In the FLC products the trails are removed by CALACS and the flux put back into the pixels where they originated.

  • zhao13zhao13 Member Posts: 8
    Thanks Jennifer, I will look at that!

  • zhao13zhao13 Member Posts: 8
    Hello Jennifer,
    Should I use the FLC files from HST archive or should I do the removal manually with "PixCteCorr"?

  • Member, Institute_Staff, Moderator_DrizzlePac Posts: 28
    We recommend that you obtain these from the archive.
  • zhao13zhao13 Member Posts: 8
    I'm asking this because the FLC files I got from the archive still have those trailed patterns. I compared FLC files with corresponding FLT files, and they seem not to have very large differences.
  • Member, Institute_Staff, Moderator_DrizzlePac Posts: 28
    edited July 6
    The CTE correction is not perfect, so you may see some residual trailing, but if you blink the FLC and the corresponding FLT, you will see the improvement, especially towards the center of the detector (eg. the bottom of the chip for extension [4] and the top of the chip for extension [1].)

    Just For reference, CTE tails only go in a single direction, whereas bleeding due to saturation will extend in both directions along the column.
Sign In or Register to comment.