Quads matched, but different quads created in reference and query catalogs
In Commit 087a515b, the matching of the X-Y quads and the RA-Dec quads have been done, and they do indeed mostly find very similar quads, for example these two:
However, even when making roughly 40,000 quads (20x20 grid, with 100 quads per grid), it couldn't find a single matching quad that actually matches with the stars! So I tried the other extreme in Commit 087a515b: 2x2 grid, with one quad per grid. You can see the resulting quads found for the reference (left) and query (right) catalogs here:
As you see, even though we are dealing with the brightest stars in the field, we are getting completely different quads. So we should find a way to more robustly identify quad vertices...
What do you think @sachinkumarsingh092?
Edited by Mohammad Akhlaghi