Optimal pixel spaces when combining different cameras


We are producing a joint HSC/VISTA imaging and photometric product. In my first experiments I used the SkyMap from obs_subaru to create a SkyMap on the same pixel grid as the HSC imaging. I did this partly because our previous pipeline with SExtracter and SWARP required using the same pixel basis. However given the VISTA native pixel scale is almost double that of HSC there will be significant issues such as correlated noise.

I now think the best approach will be to use the native pixel scale of VISTA. My hope is then I can merge the detection catalogues from HSC and VISTA and conduct measurement and forced photometry on those positions on both imaging sets with their differing native pixel scales. Is that possible or do the measurement algorithms depend on the same pixel space? Presumably the multiband deblender does for instance? I am in the process of trying both and measuring the differences but would appreciate any advice on recommended practice.

Going forward, to for instance Euclid with significantly differing resolution, this will presumably be more of a pressing issue.

Many thanks,


Since asking this question I have realised that the SkyMap is determined partly by the pixel scale. I am therefore not sure how to make a skyMap for two different instruments with different pixel scales but the same underlying tracts and patches. If anyone knows the appropriate skyMap configuration options to do that I would be grateful to know.

The pixel scale is an integral part of the tract definition. We currently have no facility to support multiple pixel scales.