Opened 19 years ago
Closed 19 years ago
#853 closed defect (fixed)
binned images and static trimsec regions
| Reported by: | eugene | Owned by: | Paul Price |
|---|---|---|---|
| Priority: | high | Milestone: | |
| Component: | ppImage | Version: | unspecified |
| Severity: | normal | Keywords: | |
| Cc: |
Description
this bug caused by many hours of pain: we have somewhat inconsistent logic in
ppImage if the config system defines a fixed region by value for TRIMSEC which
is not the 'fullarray' value of [0,0:0,0]. In pmChipMosaic, the mosaic images
are created using psImageSubset applying the calculated 'chipRegion'. BUT, if
the input images is a binned image, then the hardwired region (for the unbinned
image) is too large. the call to psImageSubset was not being tested for NULL,
and was silently passing on an empty image. I added an error check and code to
apply psRegionForImage to the chipRegion, which fixes the obvious problem. but
the underlying error is that ppImage is not correctly interpreting the
combination of trimsec from the config and the binning factor. probably it
should know that the trimsec value was hardwired and thus represents the
unbinned trimsec.
Added a mechanism to update CELL.TRIMSEC and CELL.BIASSEC with the binning.
Tested, and it seems to work.