Daniel Kustrin - 2024-10-05

Hi. I just had an interesting result from ASTAP (2024.09.25 and the prev version) when trying to solve an image - instead of close-to-linear DEC I got something a bit more cubic (i guess). RA is also affected but not as much. It doesn't happen on all the images taken during the same session and astrometry.net solves the problematic image ok.

This is the link to the astrometry.net solution: https://nova.astrometry.net/user_images/11109683#annotated. If you load the new-image.fits in ASTAP and click on Solve you will get both the "correct" astrometry.net solution and the "sort-of-cubic-in-DEC" ASTAP solution.

The image was created using python with astropy and ccdproc packages that i use in my pipeline. Both astropy and ccdproc are widely used. The dark corrected images solve ok in ASTAP and the strange solution only happens when the image is flat corrected. Other images taken during the same session (dark and flat corrected using the same dark/flats) look ok.. based on a random sample of 20. i have not checked every image as there are 531 of them so I was lucky to spot this dodgy one tbh.

If i look at images that work, the stars are pretty much in the same positions as in the image that fails. An example that solves in both ASTAP and astrometry.net can be found on: https://nova.astrometry.net/user_images/11109921#annotated.