search

quatrope / astroalign

A tool to align astronomical images based on asterism matching
MIT License
140 stars 44 forks source link

find_transform is not deterministic #49

Closed davidenitti closed 4 years ago

davidenitti commented 4 years ago

if I apply the same find_transform function multiple times I don't get the same transformation.

martinberoiz commented 4 years ago

Hi, yes, this could be possible. But the transformations should be very very similar, are you getting wildly different results?

The core of this issue is that we use RANSAC (Random Sampling Consensus) to find the best transformation, which is by random in nature. I recently came to know there is a deterministic RANSAC (see #47) that would give repeatable results. But it would take me a while to implement tbh.

davidenitti commented 4 years ago

thanks, I understand now, I was just wondering why there was this stochastic behavior, thanks for the reply.

martinberoiz commented 3 years ago

Hi, I know this is a closed issue, but you should check version 2.4 now that implements a deterministic approach to asterism matching.