Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: obs_decam
-
Labels:None
-
Story Points:5
-
Epic Link:
-
Team:Data Facility
Description
Pixels near the edges of the DECam CCDs are bigger/brighter and correcting them is not trivial. One way to move forward is to mask them out.
DESDM and CP mask 15 pixels on each edge. The cut was later raised to 25 pixels, with the inner 10 pixels flagged as SUSPECT.
Attachments
Activity
Code changes are merged in DM-4559.
For now I set the default config of isr.numEdgeSuspect to 35. That is 10 pixels more than the DESDM.
Hsin-Fang Chiang
added a comment - Code changes are merged in DM-4559 .
For now I set the default config of isr.numEdgeSuspect to 35. That is 10 pixels more than the DESDM. A script is also added to obs_decam/example to output the defects in text-based format. The script may be a starting point if in the future we decide to refactor and persist defects in text-based format, Footprints, etc in obs_decam (also see DM-4559).
Hsin-Fang Chiang
added a comment - - edited A script is also added to obs_decam/example to output the defects in text-based format. The script may be a starting point if in the future we decide to refactor and persist defects in text-based format, Footprints, etc in obs_decam (also see DM-4559 ). 
My original thought was to just set masks for the edge pixels in the ISR processing, so those pixels can be easily excluded in later processing. In the verification datasets meeting yesterday it was discussed that it would be good to handle them like defects.