Quadrant filters should not be used for HAP catalogs

[stscijgbot-hstdp]

Issue HLA-1551 was created on JIRA by Rick White:

The quadrant filters on WFC3/UVIS (names starting with fq) should not be used for HAP-SVM catalog generation. Two changes are needed in the logic:

Do not include fq images in the total image.

Do not make catalogs for the fq images.

The fq filters have different bandpasses in different parts of the image. That means that the photometry and detection thresholds need to be different in the different quadrants. That would require major changes in the cataloging software, which are too complex to consider in the HAP processing.

The correct thing to do is simply not to use the fq filters for the catalogs. That means they should also be omitted from the detection image. Currently the fq filters apparently are included in the total image, and catalogs are created. This should be fixed.

An example is visit hst_14393_18_wfc3_uvis_id1x18, which includes two filters, f606w and fq492n. In the current DrizzlePac 3.10.0 processing, that visit has both catalogs rejected with zero sources. The rejection is very likely due to inclusion of the fq492n image in the total image. The differences in the 4 quadrants is clearly visible in the total image.

!hst_14393_18_wfc3_uvis_total_id1x18.jpg|width=512!

[stscijgbot-hstdp]

Comment by Rick White on JIRA:

I assigned this to Steve Goldman but don't know who might work on it (if anyone does). I believe the change is probably pretty simple: there are already other filters (such as ramp filters and polarizers) that are excluded from the detection image and do not have catalogs. It should be possible to add these quad filters to that code.

[stscijgbot-hstdp]

Comment by Steve Goldman on JIRA:

to be closed by #⁠2093.