[daip] question about FLATN, OHGEO

[Deborah Haarsma]

Sorry for being impatient (we have our own local deadline on this project).  We very much appreciate you taking the time right away to look at the files and reply - this gives us some good options to pursue. 

Again, thank you very much - 

Deb Haarsma


>>> Eric Greisen <egreisen at nrao.edu> 8/11/2009 11:05 AM >>>
Deborah Haarsma wrote:
> Dear analysts,

daip - not analysts

> 
> Larry Molnar and I are long time radio astronomers, but we also work with data from Calvin College's small optical telescope.   Amateur astronomy software works well enough for basic optical reduction, but when we need a specialized task we tend to turn to AIPS.   One of our students, Jess Vriesema, has photographed 98 CCD images in the field of Andromeda and we want to combine them into a large mosaic.   He ran the individual images through Pinpoint(tm), an astrometry engine which we have found to work well in other projects.  Then we brought the images into aips to correct the geometric projection over the large field.  The FLATN procedure seems ideal for the task; when we run it, the global geometry seems corrected, but there are double stars on the boundaries between fields.   We then ran OHGEO on the files individually to get a better feel for what's going on, referencing them to the center of Andromeda a few fields away.  I've attached four files - two different field
 
s (S32 and S33), before and after OHGEO.   The two fields have a strip of overlap (S32 is south of S33); the stars in this strip have the same RA and Dec before OHGEO, but rather different RA and Dec afterward.   In the headers of the files after OHGEO, you can see that AIPS has moved the CD and TR keywords into the history.  I expected OHGEO to keep modified versions of these keywords in the header.  Do FLATN and OHGEO read the CD and TR keywords?  Is there some option we're setting incorrectly in order to do the mosaic?  We appreciate your help on this matter.  We can send further data files as you have questions. 

You are being impatient to expect a reply in less than 1 day.  But I 
have taken a look and it is what I was afraid of.  AIPS has not been 
updated to understand the "new" WCS conventions - remember aips was 
supposed to be replaced in 1992 (!) so long-term future things have not 
been done as much as they should have been.  In particular AIPS does not 
fully understand the CD matrix and I have not heard of the TR 
parameters.  They may be in Paper IV which has never been fully worked 
on yet say approved but they are not anywhere else.  I note that aips 
FITLD does know enough to convert the CD matrix into rotations and 
slightly different axis increments.  But TR is ignored and put in the HI 
file at that point.

You need to find a program that will re-grid your optical data so that 
the TR are not needed.  It would be good to eliminate the CD at the same 
time.  Then FLATN could do its thing.

I know that ds9 understands the full WCS - does it have a regrid and 
output option?  Perhaps casa (former aips++)?

Mark Callabreta wrote wcslib and may be able to help you (atnf.csiro.au)

Eric Greisen