[daip] coordinates in AIPS

[Eric Greisen]

Anna Bartkiewicz wrote:
> Dear Eric,
>   Thank you very much for your explanations.
> Is that possible, you could send me the 
> formula of CD1_1, CD1_2, CD2_1, CD2_2 or where I could find it?
> Is that possible that I could fix 
> the rotation as 0 (CD1_2 and CD2_1=0???)
> and leave CD1_1 and CD2_2 as 0.0219"?

Perhaps you have not understood.

Someone has run astrometric software on your image and fit parameters to 
the coordinates.  The CDi_j are the result of that fit - along with the 
CRPIXi, CRVALi, CTYPEi, etc.  The pixels in your image are not on a 
"simple" projective geometry, although they are not terribly far from 
one (the usual optical telescope -TAN).  They are warped so as to be on 
a rectangular grid but even that is twisted.  The amount of the warping
becomes several pixels by the edges of your image.

There is a software library WCSLIB that can handle this complexity.
Unfortunately, AIPS has never been overhauled to replace all the careful 
coordinate work we invented ahead of WCSLIB to use the more advanced 
library.  Other software has been revised to use WCSLIB - public 
packages such as ds9 for example or iraf.  Try using one of these to 
read in your image - perhaps they have an image regridding routine that 
will take your skewed image and recompute it onto a more regular 
projective grid.  AIPS has such a task (OHGEO, HGEOM) but it is not 
smart enough to handle your input image.

Eric Greisen