Re: Star's magnitude adjustments

[Chris Albertson <chris@topdog.pas1.logicon.com>]

Mike,

I thought of a few ways to test your program.

1) If the gradient is caused by errors in the flat all images
processed with the same flat should produce about the same
correction vector.  Any change should be smooth over a period
of days

2) Try introducing error into the flat and reprocessing.  Add
a gradient to a flat and see if the gradient appears in the
correction vector.  This should provide a very powerful "proof"
that your techique does fix bad flats

3) Build a few "fake" star lists.  don't use TASS data.  Maybe
GSC reformatted to look like star.exe's output file.  Your
correction vector should be real small and random.

The possible error that I think your system could introduce
is something akin to a servo with the gain set to high.
You likely need some kind of low pass filter on the rate of
change on the correction vector.

I still think your system is best used to provide feedback to 
the flat generation process.  The correction vector could be
used to fix the flat and the data is reprocessed.  Repeat this 
untill you get a "small" correction vector.

The best solution is still building a <<1.0% flat field device.
Should not be to hard with our 52 mm diameter lens
-- 
--Chris Albertson

  chris@topdog.logicon.com                Voice:  626-351-0089  X127
  Logicon RDA, Pasadena California          Fax:  626-351-0699