Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: ip_isr
-
Labels:None
-
Story Points:2
-
Epic Link:
-
Team:Data Release Production
-
Urgent?:No
Description
The `Exposure.getReadoutCorner()` method used to determine the appropriate X/Y flips needed to properly orient two amplifiers to match returns the incorrect value for some amplifiers of the LSST CCD.
Ex: Both Segment C10 and Segment C00, which are on different sides of the CCD midline break return the same location for readout corner:
```
detector = exp.getDetector()
amp_top = detector['C10']
amp_bottom = detector['C00']
print(amp_top.getReadoutCorner())
print(amp_bottom.getReadoutCorner())
> ReadoutCorner.LL
> ReadoutCorner.LL
```
This is an issue for calculating crosstalk between amplifiers on opposite sides of the midline break as the orientation of the two pixel arrays obtained using `lsst.ip.isr.CrosstalkCalib.extractAmp()` is incorrect (see attached images).
Attachments
Activity
Jenkins success:
https://ci.lsst.codes/job/stack-os-matrix/35741/display/redirect
I've also attached the notebook for my check. ct-readoutCorners.ipynb
I believe I've tracked this to an issue in afw. After doing all of the appropriate shifts and flips, the rawFlipX/Y parameters were set to False (as the flips had been done), but the readoutCorner was not. I've added the appropriate changes to that now, and the attached image shows that the results of CrosstalkCalib.extractAmp now look correct.