Difference between code-tolerance and pixel-error

[Robert Siverd]

Hi Dustin,

Can you briefly explain which aspects of the solution are controlled by the --code-tolerance and --pixel-error options?

I once thought they might be used for different procedures (i.e., code-tolerance for star identification and quad matching, then pixel-error for polynomial tuning) but reality seems more complicated.  Testing shows that both affect which quads match and the logodds value before the tweak process even begins. I'd really appreciate a quick clarification.

Thanks,
Rob

[Toni Šarić]

Raising from the dead.
I am also curious about those two parameters.

[Dustin Lang]

--code-tolerance is about quadrangle matching and determines which possible matches will be evaluated.

--pixel-error is used when evaluating potential matches, and also during tweak.  It is supposed to be the uncertainty on detected star positions.

So yes, both would affect which matches will be accepted.

--
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/6f9482e4-56ef-46c7-8133-e9bdc63fc333n%40googlegroups.com.