RA&DEC residuals

[Ning B]

I have some images ,FOV is 2.3°×1.5°

I use wcs-xy2rd to get the ra,dec (SIP=3),and crossmatch with gaiadr2 catalog 

here is my residuals picture ,I can't explain this problem about ra's residual

thanks for your any help

 

[Dustin Lang]

Hi,

You could try creating a "quiver plot" -- plot in RA,Dec or in pixel coordinates, and at the position of each matched star, plot an arrow showing the delta-RA,delta-Dec.

If the SIP solution is wrong, you should see nearby arrows all pointing in the same direction, and then the direction slowly changing as you move around the image.

cheers,

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/cc7309b4-9e6d-4ef4-969e-be1b85d15613n%40googlegroups.com.

[tano...@gmail.com]

Hi Dustin, hi everyone

I recall an old issue about the ra dec residuals.
I have a question. Why the arrow should pointing in the same direction if the SIP solution is wrong?

Cheers

Gaetano

[Dustin Lang]

Hi,

Have a look at the attached plot.  See that the matched quadrangle of stars is in the top-left corner, and the SIP solution is not good.  The further away from the matched quad, the worse the matches are, and if you looked at the distance from the red circles to the green circles, in a little area of the image, those vectors are all in the same direction.

Hope that helps,

cheers,

dustin

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/185134bd-cda1-44cb-87d2-aee404b16c7an%40googlegroups.com.

[Gaetano Zarcone]

Ok.

I think I got it, but just for a confirm from you: by seeing the attached plot, would you say that the SIP solution is good?. 

This is the result of a field solving of a wide fov with tweak order 5

cheers

Gaetano

[Dustin Lang]

Hi,

To my eye, I don't see any coherent shifts.  With a good 5th-order SIP fit, the residuals would be 6th order and higher.

In that plot, did you make all the vectors the same length?

cheers,

dustin

[Gaetano Zarcone]

Hi Dustin,

yes the vectors have same length. 

I update the plot, see the one attached. here each vecto is proportional to the ra dec residuals (in arcseconds) and multiplied by 50 in order to have a better visualization

what do you think?

cheers,

Gaetano

[Dustin Lang]

That looks fine -- no large-scale coherent patterns that I can see!

[Gaetano Zarcone]

Ok

Thanks Dustin

Cheers

Gaetano