find wcs when pointing unknown
Guest wrote on Jun 03, 2006
Is there any IRAF task(or non-IRAF for that matter) that will take an image with no wcs header info and locate the sky postion and produce a wcs? I know the position to about +/- 10 deg in each axis.
Thanks,
Jim
Thanks,
Jim
Mike Fitzpatrick wrote on Jun 03, 2006
Depends, but probably not considering the error is 100 deg^2. How big is the field of view? How deep is the image?
The WCSFIXER task (search the site here for information or http://nvo.noao.edu/wcsfixer for a direct link) is meant to do this but needs at least an approximate (+/- half the field width) center and scale. If your FOV is 5 deg or so the algorithm would work but the current reference catalogs would be much too dense. You could use something like the Yale Bright Star Catalog to identify the field a bit better and manually enter that position, but again it depends on the depth of the field as to whether you have enough stars to match.
On the non-IRAF side, the IMWCS task in the WCSTOOLS package from SAO does something similar although I've never used it. Google for a link and similar results or see http://tdc-www.harvard.edu/wcstools/.
The idea is either case is the same -- match a reference catalog to a catalog of objects created from the image, compute a plate solution and update the header. There are tools to help with these steps, but the closer the image catalog is to the portion of the reference catalog you select the better luck you'll have. Matching 100 image stars to 1,000,000 catalog stars produces too many false solutions (and takes forever to run). Hope this helps.
Cheers,
-Mike
The WCSFIXER task (search the site here for information or http://nvo.noao.edu/wcsfixer for a direct link) is meant to do this but needs at least an approximate (+/- half the field width) center and scale. If your FOV is 5 deg or so the algorithm would work but the current reference catalogs would be much too dense. You could use something like the Yale Bright Star Catalog to identify the field a bit better and manually enter that position, but again it depends on the depth of the field as to whether you have enough stars to match.
On the non-IRAF side, the IMWCS task in the WCSTOOLS package from SAO does something similar although I've never used it. Google for a link and similar results or see http://tdc-www.harvard.edu/wcstools/.
The idea is either case is the same -- match a reference catalog to a catalog of objects created from the image, compute a plate solution and update the header. There are tools to help with these steps, but the closer the image catalog is to the portion of the reference catalog you select the better luck you'll have. Matching 100 image stars to 1,000,000 catalog stars produces too many false solutions (and takes forever to run). Hope this helps.
Cheers,
-Mike
Emilliano Gregori wrote on Jun 03, 2006
Hello,
maybe this could be useful:
"Match: a program for matching star lists"
http://spiff.rit.edu/match/match-0.8/match.html
I did not try, but I read somewhere that it's used in some ACS pipeline, or something similar... I do not remember...
Jim, if you are successful with "match", please tell me, I'm also interested in it for an automated reduction pipeline...
Cheers,
Emiliano
P.S: and there is also SCAMP:
http://terapix.iap.fr/soft/scamp
from the French TERAPIX project...
maybe this could be useful:
"Match: a program for matching star lists"
http://spiff.rit.edu/match/match-0.8/match.html
I did not try, but I read somewhere that it's used in some ACS pipeline, or something similar... I do not remember...
Jim, if you are successful with "match", please tell me, I'm also interested in it for an automated reduction pipeline...
Cheers,
Emiliano
P.S: and there is also SCAMP:
http://terapix.iap.fr/soft/scamp
from the French TERAPIX project...
Guest wrote on Jun 03, 2006
Well, the FOV is about 15 min sq. Not very big. I had looked at these tasks and they all seem to need some coords that are within a few min of being correct. No way I can do that, for a single pass. Now I could break it down into many small passes, each using a different set of coords, and cover the entire area in 100, 200? some large number of runs. Not something I would look forward doing.
Thanks for the ideas,
Jim
Thanks for the ideas,
Jim
Last post on Jun 03, 2006