shift() not working

[lukfd]

Hello,

When I try to run chi2_shift() and the shift.shiftnd() the resulting picture will not have any shifting applied to the rgb.

The code that I wrote is, in which variables r,g,b are the np.array got from r = fits.open("data/"+"M66-Red.fts")[0].data :

image = r
shifted = g
dx,dy,edx,edy = image_registration.chi2_shift(image, shifted, upsample_factor='auto')
g = shift.shiftnd(shifted, (dx, dy))

image = r
shifted = b
dx,dy,edx,edy = image_registration.chi2_shift(image, shifted, upsample_factor='auto')
b = shift.shiftnd(shifted, (dx, dy))

and on the console I am getting a Future warning:

C:\Users\Luca\Git\ustobservatory\newscripts\image_registration\image_registration\fft_tools\zoom.py:101: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  outarr[ii] = outarr_d[dims]
C:\Users\Luca\Git\ustobservatory\newscripts\image_registration\image_registration\fft_tools\shift.py:103: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  for dim,(off,nx) in enumerate(zip(offset,data.shape))],

Am I using the methods in a wrong way? Should I use another method to make the shift of the images?

This is the final image, as you see the

Thank you so much! Really appreciate all the work !

[keflavich]

Very cool image! At a glance, though, the problem is almost certainly that your image is totally dominated by the sky background. The galaxies and stars are only a tiny perturbation on that huge sky emission. You should try to remove the background first - for the image_registration algorithm to work, it needs the background to be close to zero on average.

If you can't get rid of the background by flat fielding, a good thing to try with an image like this might be to perform a really big median filter, then subtract that from your original image, then do the shift measurement. You can then take the measured shift and apply it to the original photo.

[keflavich]

(for my reference) See also https://stackoverflow.com/questions/61449527/image-registration-python-package