search

spacetelescope / drizzlepac

AstroDrizzle for HST images.
https://drizzlepac.readthedocs.io
BSD 3-Clause "New" or "Revised" License
52 stars 38 forks source link

Implement SVM image comparison with HLA products #560

Open stsci-hack opened 4 years ago

stsci-hack commented 4 years ago

Standard pipeline processing includes (as of HSTDP 2019.5) comparisons of drizzle products to flag obvious errors in the alignment that would result in smeared or double exposures. The comparisons implemented for that processing would be suitable for use in comparing the HLA-generated products with the new single visit mosaic (SVM) products. Work should be done to see what comparisons would make the most sense for providing some indication of how well the SVM products compare to the HLA products.

Part of the difficulty in making this comparison comes down to the final shape/size of the drizzle products being (slightly?) different. It may be sufficient to say that for the same part of the sky (mosaic) the 2 mosaics are similar enough to not impact any subsequent data analysis.

stsci-hack commented 4 years ago

Another issue to contend with when trying to compare the results: putting the HLA results on the same WCS as the HAP results (HSC vs GAIA2?). This becomes critical to being able to identify what the HLA thinks is the same part of the sky as the HAP in order to compare and see whether they are similar.

It may end up just being easier to try to measure the focus instead, if it can be turned into a value that is easily understood and relatable back to the data (such as FWHM).

stsci-hack commented 4 years ago

Another complication comes from the fact that HLA uses a plate-scale of 0.09(?!?) arcsec/pix for their WFC3/IR products, compared to the HAP 0.128 arcsec/pix. This makes it impossible to do any comparison of image quality between the products.