Suspense's profile - activity

LBerger, that sounds like the kind of thing I'm looking for. How reliable will the sorting and matching steps be considering that these photos can be very noisy?

Thanks for the suggestions, Eduardo. But I'm not interested in using existing software. My purpose is learning ways to do this myself. See my edit for more information.

I'm looking for a way to align astrophotos with OpenCV in Java. These photos will typically have a dark background, random scattering of stars, and a subject like a galaxy or nebula. They will also have high levels of noise. I could come up with an algorithm myself, but there must be something I can use that already exists.

Right now I'm successfully using Video.findTransformEcc() to align images that are already very close. However, I also have photos like the following examples I'd like to align. Can anyone give suggestions?

EDIT for more information: My goal is to sample pixels from each image in the stack based on its location in the sky, rather than its location in each image. This will let me analyze the variance in that location to filter out noise and bring out faint details. I'll go with whatever technique works best, whether that be alignment or stitching or something else.

Also, I'm not interested in finding existing software for this. I'm aware of several existing software packages for this exact use-case. My goal here is to learn how to do it myself.