Post flat processing

["Gutzwiller, Michael" <mgutzwiller@lanvision.com>]

I've slowly been creating a program to process star lists to remove the
flat field errors.  I've converged on the following solution and would
like to share it with everybody for consideration and helpful criticism.

The approach is as follows:

*	An entire night's run for each camera is examined separately.
*	Each star list for a camera is checked against stringent
requirements for photometric quality.  This ensures that the matching is
done with good TASS data.
*	For good lists the stars are matched with high quality stars in
the Tycho catalog.  High quality stars have sigV and sigB less than
0.060 and are separated from other stars.
*	The matched stars are added to a list containing the declination
and the difference between the Tycho magnitude (Arne's transformation)
and the measured magnitude.
*	After all lists for the camera have been processed each camera
is divided into (currently 4) equal regions of declination.
*	A median line fit of delta magnitude vs. declination is made for
each region based on the stars in that region and halfway into the next.
The overlap ensures that the endpoints match better.  The median line
fit is less sensitive to outliers.
*	The median line fit can then be used to adjust the TASS
magnitude for all stars in all lists based on its region.

	There are a few unresolved issues:

*	Is four regions enough?  Test data so far says ok.
*	How should the adjustment be documented?  Maybe new list shows
adjustment taken for each star?
*	Should color transforms be computed?  The data is there and the
computation would be easy to add to the list header.


Comments?

Mike Gutzwiller