StarAlignment registers these images without problems with default parameters.
I also dont understand why limiting the max number of stars to 500 breaks the matching of stars when only 30 or so stars are actually found. Counter-intuitive. Maybe someone who is in the know can comment.
When the
maximum stars parameter is set to a nonzero value, SA's automatic probing routine is disabled. Since these images are so "soft" and noisy (especially the unbinned image), SA detects too many false star structures in its first attempt, and fails. A second attempt with more restrictive parameters succeeds, but for this to work automatically,
maximum stars must be set to zero to enable the automatic probing routine. In the BPP script, this parameter is set to 500 by default because this greatly speeds up the preprocessing task and works fine in most practical cases, where automatic probing is not required.
I implemented a noise reduction parameter to help fixing these issues. By setting
noise reduction = 8 in the
Registration parameters section of BPP, these images can be aligned without problems with the default 500 stars limit. Setting maximum stars to zero also works in this case, but the noise reduction parameter is a more specific solution.
