Astrometry issue for some images (SOFI and EFOSC)

[thespacedoctor]

Here's a reduced SOFI image for SN2012ec overlayed with 2MASS sources. There are a few isolated point sources in the field but the astrometry solution is not perfect:

I know SN2012ec is in a particularly difficult region to get a solution for due the the large number of sources identified in the galaxy.

[thespacedoctor]

Also see this image for EFOSC (LSQ12btw) - some of the stars are off by a few pixels (ESO will not accept this image):

[thespacedoctor]

And again for EFOSC (pg0231_20121212_i705_56972_1_fr.fits):

[svalenti]

solve astrometry for small EFOSC and SOFI images is complicate. we have a rate of > 90% of success. For the other images, the user/reducer can use PESSTOASTRO in interactive mode

[thespacedoctor]

Reopening - someone need to add this to the Pipeline user's manual

[thespacedoctor]

Cosimo could you add this item to the user manual whenever you test PESSTOASTRO please?

[cinserra]

I tried PESSTOASTRO -i -a but it failed. I am assuming because there weren't enough stars in common between the blue and the red. Is that correct? However what catalogues are those, especially the red that has such few stars (at least in this image)?

Furthermore, the red catalogue has correctly identified only 1 star on the entire frame, while others are not visible( n1 and n8 for example)

[cinserra]

Same problem with pg0231. If the pipeline does find three stars present in both catalogues, it does the astrometry but it is totally out.

[svalenti]

some fields can be particularly tricky. From the image, I clear see that the initial position are wrong. One can try using options to give a different starting position to compute astrometry --crpix1=CRPIX1 crpix1 [] --crpix2=CRPIX2 crpix2 []

the option -c CATALOGUE allow to use a different catalog or even an external catalog

[svalenti]

I tried to do astrometry and i did not have problem. PESSTROASTRO in interactive mode is asking to select one star (better to select one isolated star). This is needed to understand the shift between catalog and detection. If the shift is too large use option --crpix1 and --crpix2. Once the initial shift is correct, the script ask for 3 stars to compute astrometry. I don't think this is a bug, but more that it is not clear how to use the script in interactive mode

[cinserra]

OK, it is not needed soon, but Stefano, can you put a screenshot of your middle stage of the procedure (like those reported above) and final astrometry, so I can duplicate your findings and write something in the wiki/user manual? There is no rush in doing it.

[svalenti]

running PESSTOASTRO pg0231_20121212_i705_56992_1_fr.fits -i -a I get this screen with sextractor objects in blue and catalog in green there is a clear shift between blue and green. First request from script is to select one star in the image. I select an isolated star (number 7 blue) and give the corresponding number of the green circle (12) if the source (blue) and the shifted catalog (red) are now close enough, select three stars in the image and the astrometry will be done