Deconvolution and multisession images

[Ignacio]

Hi,

I am processing an image done in two different days, with quite different conditions. Most notably, seeing. I have calibrated and stacked each session independently, and  now face two different paths. 1. Combine the two master lights (say, to maximize SNR) and proceed with deconvolution, or 2. deconvolve each individual master with its specific PSF, and then combine them.

To put it in context, in one session the average FWHM is about 3", while in the other is a bit over 4". Both have about the same SNR.

I can always try both paths, but I am trying to reason if in theory there is a better one. My guess is to do independent deconvolutions, trying to match the final FWHM, and then coadd them.

Thanks,
Ignacio

[Philip de Louraille]

It will be interesting to see the differences between both results. Decon seems to bring tiny artifacts (especially near stars) so combining 2 different decons and then doing a histogram expansion might show some curious results near stars.

[Ignacio]

After playing with this a little bit, I ended up applying 2 deconv iterations to the latter master light, in order to bring the FWHM to comparable levels. Then, I combined both master lights and applied 80 deconv iterations to the final stack. Results were somewhat better in terms of color halos, compared to not equalizing the FWHMs. Yet I wonder if there is a better method.

A FWHM limit stop in the deconv tool would help in this process.

If you care to see the final image: http://astrob.in/full/35216/B

Ignacio