I have been experimenting with using generated star fields as the reference for registering sets of subs.
This usually works fine, but occasionally I get errors in finding sufficient numbers of stars or matches. I've found a case that seems particularly interesting, and I cannot figure out what's going on. I have generated the attached (zipped) FITS file with a generated star field. When I run StarAlignment in Detected Stars mode, it does not find any stars in this image.
I have experimented with all of the detection parameters in StarAlignment without success. Sometimes it will detect a single star, but most of the time, it detects no stars at all. I ran the star detection on actual subs of this approximate field, and it found stars with no problems. If I blink the subs and the synthetic star field, it's visibly obvious that they are the same field.
In a fit of desperation, I thought about any differences between the actual subs and the generated field. On a whim, I decided to add some noise to the generated image. I used NoiseGenerator to add 0.0001 data numbers of uniform noise, and then StarAlignment detected all of the stars with no problems. All of the subs registered to the generated field fine after that.
Do you have any ideas why no stars are detected with a perfectly black background?
Thanks,
-Wade
zip
zip
Comments