Looks great Martin and you're on the right track.
Ivo's comments go to the heart of processing and why stars matter. This is exactly how it's done.
Quote:
Originally Posted by irwjager
Stars should readily show their diffraction patterns and definitely not look like blown out fuzzy blobs, smudges or featureless discs. Their over-exposed cores should not grow or bleed into neighbouring pixels when being stretched from their linear state and should always be well defined. Just like black clipping is a no-no to many of our peers, so is white clipping.
Diffraction patterns serve a few purposes to the viewer and person processing the data;
They serve as a good indicator as to what instrument was used, they serve as color balancing tool and star temperature indicator (in case of a visual spectrum dataset), they serve as a deconvolution calibration target and they serve as a proof/authentication of detail fidelity. The latter two are because they always constitute high SNR areas in your image (even in poor SNR datasets!). This makes it easy for deconvolution to do its thing and thus provide an ideal target to gauge how well the algorithm is doing. And due to to the intricacies of their patterns, they show up any artifacts much easier as well, giving a more experienced viewer a good idea of how an image was processed and where it might fall on the astro "art" vs documentary astrpphotography continuum if that is a concern.
|