RE: Bad pixel determination?

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

While the ADC may run in two's complement mode the FITS headers have a
BSCALE of 1.0 and a BZERO of 32768.0 which scales the pixels to run from 0
to 65535.

I have analyzed some of the dark files from the 1836 series.  A median of 5
dark images for the V filtered camera shows a range of 3629 to 20756 with a
median of 3677.  There are some hot pixels but there are no completely
invalid pixels in the imaging area from what I'm seeing.

There do seem to be quite a few cosmic rays in any one image which are quite
visible in the dark images.  That's why I use median darks for subtraction
in image processing.


Thanks,

Mike G.

-----Original Message-----
From: aah@nofs.navy.mil [mailto:aah@nofs.navy.mil]
Sent: Thursday, December 07, 2000 10:58 AM
To: tass@listserv.wwa.com
Subject: RE: Bad pixel determination?


Rob Creager should look at the TN's if he
hasn't already; I think Michael did a check of the
dark current when Mark IV images first started
being available.  There should be some ideas in
there as to which pixels to use and how to set
the bins for the histogram.
  Tom is running the ADC in two's complement
mode, so the numbers really do run from -32768 to 32767,
and not 0-65535.  Be sure you are doing the conversion right.
  You should use two images for checking for bad pixels.
The dark current image tells you whether a given pixel
has a strange dark current.  A flatfield (say, twilight image)
will tell you about the response to light, identifying
those pixels with no response or abnormal response.
Arne