[fitsbits] Plate Center on POSS2 Plates

[Doug Mink]

> I've just retrieved a fits file from
> http://archive.stsci.edu/cgi-bin/dss_form
> and would like to know the coordinates which the center pixel of this
> image corresponds to.
> I notice that there are comments "Plate center RA" and "plate
> center DEC" next to keywords PLTRAH,PLTRAM,PLTRAS etc. but for the most
> part, these seem not to correspond to the center of the actual fits image
> since they are ~ few minutes different from
> the epoch 2000 coordinates and the star of interest lies fairly close to
> the center (~ tens of pixels from center). Can someone explain this to me?

The PLT plate center in the FITS header is the center of the photgraphic plate
which was scanned to create the image from which your image was extracted.
There are other parameters which record the offset of your image from that
center.  My WCSTools package, at http://tdc-www.harvard.edu/software/wcstools ,
can decode the DSS headers (as well as FITS standard WCS) and return the sky
position of any pixel.  There is a program in the package, imsize, which
returns the coordinates of the center of the FITS image based on the world
coordinate system defined in its header.  The xy2sky and sky2xy convert between
image pixels and sky coordinates.

-Doug Mink
  Telescope Data Center
  Smithsonian Astrophysical Observatory
  Cambridge, Massachusetts USA