Code review of "corrected" star positions at acquisition
Closed this issue · 2 comments
We should do a review of the code that determines the "corrected" star positions at acquisition to confirm correct/continued use for predicting that a star "would have been identified" in the context of the multiple star flag.
Question: Are the attitude quaternions being applied properly to determine expected star positions?
Answer: Looks good. JMC was a bit confused about when aca_misalign is applied, but the code is correct (even if it uses an unnecessary identity matrix for aca_misalign) and confusion has been alleviated. https://github.com/sot/mica/blob/master/mica/stats/acq_stats.py#L121
Question: Is the process to check against proper-motion corrected AGASC position after the one-shot has been applied good enough to use for these post-facto position checks for the "would the star have been identified" idea? On-board it is using the distance from the tracked star to each other star.
Answer: Yes.
Question: Is the (img_func == 'star') check sufficient in the "would the star have been identified" code?
Answer: There is a mismatch here between the limit of the corrected position offset (5 arcsec ~ dr) and the on-board id dist limit. Probably fine.
Looks good!