Matching stars output

Toni Šarić

unread,

Feb 22, 2021, 6:39:03 AM2/22/21

to astrometry

If I understood output of astrometry correctly:

simplexy: found 163 sources. Solving...

... log-odds ratio 30.4066 (1.60485e+13), 10 match, 0 conflict, 33 distractors, 56 index.

163 is number of possible stars

10 is number of matched(index connected with source) stars

56 is number of stars in the index.

Then, why does created .corr fits contains 44 rows(in this case) i.e. 44 matched stars?

Dustin Lang

unread,

Feb 22, 2021, 8:16:21 AM2/22/21

to Toni Šarić, astrometry

Please send an example; it's difficult to speculate without seeing what you're seeing.

--
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/a0d97f8d-da6f-419e-bbde-c96b49dfeacen%40googlegroups.com.

Toni Šarić

unread,

Feb 22, 2021, 8:42:27 AM2/22/21

to astrometry

Sorry, my bad.

Here is the output, because I got an error when tried to upload .corr or .fits file.

Screenshot_2021-02-22 Untitled - Jupyter Notebook.png

Screenshot_2021-02-22 Untitled - Jupyter Notebook2.png

BR

Toni

Dustin Lang

unread,

Feb 22, 2021, 9:37:52 AM2/22/21

to Toni Šarić, astrometry

It looks like those reported number of hits, distractors, etc are only including field stars up to the *best* solution found, while the corr file also includes additional matches after the best one.

What happens is: given a match, we predict where image stars will be found, based on where index stars should be. When an image star matches, the logodds go up. If an image star is not matched, the logodds go down, so the total log-odds increases or decreases after looking at each image star. It is kept or rejected based on the best log-odds found, but the corr file includes all matches.

--dstn

--
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/e89c5988-a38f-4e51-8771-90bf1b8a46b3n%40googlegroups.com.

Toni Šarić

unread,

Feb 24, 2021, 4:38:42 AM2/24/21

to astrometry

So, you are saying this output in the terminal is not the final result. More stars are matched later as can be seen in .corr.
Also, (RA, Dec) coordinates of the center of the field are not the same as from the header.

Toni Šarić

unread,

Mar 5, 2021, 3:04:23 AM3/5/21

to astrometry

Hi Dustin,

I noticed another different output or outputs.
Not just the RA, Dec(CRVAL1, CRVAL2) are different in output and header, but scale(arcsec/pix) also.
Which result I should take as the final accurate one?

BR,

Toni

Dustin Lang

unread,

Mar 5, 2021, 7:41:16 AM3/5/21

to Toni Šarić, astrometry

Not sure what you're talking about.

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/e4ef71ea-185d-4f76-8e76-10a592165c80n%40googlegroups.com.

Toni Šarić

unread,

Mar 5, 2021, 8:14:31 AM3/5/21

to astrometry

I can not append output to a chat so I screenshoted output from terminal and header from a new fits.

You can see that RA, Dec are different but also arcsec/pix scale.

1.png

2.png

3.png

Dustin Lang

unread,

Mar 5, 2021, 8:42:15 AM3/5/21

to Toni Šarić, astrometry

The first one is the initial match, before any tuning up.

You are probably setting --crpix-x, --crpix-y, right? The CRPIX values correspond to the reference point CRVAL, which is not necessarily in the center of the image. In your case, those RA,Dec values differ by half a pixel.

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/63bf4214-9f41-43db-884b-811e602674e1n%40googlegroups.com.

Reply all

Reply to author

Forward