Force solver for using a minimum number of stars?

[andreas....@gmail.com]

Hi,

is there a way to force the solver to have a minimum number of stars?

I noticed that for seemingly similar frames I sometimes get ~20 stars and then sometimes ~30. This sometimes also correlates with the result to be different in a way to have the 2d-polinomial to shift the ra/dec values of the pixels by several 10 pixels.

So for two things it would be interesting for me to set the minimum number of stars.

First, to not have this big ra/dec changes on pixel levels. I project the satellite paths on it based on ra/dec and that shifts the satellites too much to be usefull.

And secondly, more stars would allow me to use more stars and their brightnesses for my satellite brightness comparison.

I checked --help but non of the options seems to be what I am looking for.

Can someone maybe help?

Best regards,

Andreas

[Dustin Lang]

Hi,

Probably what is happening is that it's finding a solution using a different index file.

You might see it printing at the end something like "found solution in index-5207-08" or something like that.  I bet you'll find that the index scale numbers are different when you get 20 vs 30 stars.

You *could* edit your astrometry.cfg file to only use one of the index files.  That's messy though!

One thing you can try would be to re-run the solve-field program, and pass in the WCS found the first time around using

    solve-field --verify IMAGE.wcs [and then the usual arguments you use]

That should check the solution in all available index files and keep the best one.

Another thing you can do to make the solutions more stable is to set "--crpix-center".

I hope that helps!

-dustin

--
You received this message because you are subscribed to the Google Groups "astrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to astrometry+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/459ded13-66d0-44ea-b5a7-4e221ab26951n%40googlegroups.com.