Interesting...
My application was written for SiLab processor F330, it was developed to track the laser beam (for drilling straight holes in the ground), so the task was very similar to astroguiding, and based on what was mentioned on CN discussion, the algorithms used are similar - I also first found a square of n-pixels where the bright spot was located and then generated voltage for hydraulics based on spot position inside that square (to speed up the process.. the scan of the whole frame took 1.5 seconds due to processor speed).
Maybe I will try again one day, when/if time permits.. there are so many things I would like to finish before I kick the bucket
EDIT: I just remembered I wrote about it on
this forum 12 years ago.. how time flies, it is unbelievable.