RE: Bad pixel determination?

["Albertson, Chris" <CAlbertson@primeadvantage.com>]

When you computed the histogram did you look at _all_ of the
pixels in the file or just those in the image section?

The image pixels are surrounded by non-image pixels.  So it
is not surprising that if your histogram looked at non-image
pixels to find many "bad" pixels.

I don't remember off hand if Tom's header files include an
"IMAGESEC" (sp?) keyword.  


> -----Original Message-----
> From: Creager, Robert S [mailto:CreagRS@LOUISVILLE.STORTEK.COM]
> Sent: Wednesday, December 06, 2000 3:01 PM
> To: 'Tass Mailing List'
> Subject: Bad pixel determination?
> 
> 
> 
> I was just starting work on some code to determine bad pixels, and was
> wondering if anyone has done this or has comments?
> 
> I'm looking at the dark frames for V/I band, and they look 
> very different
> (this is data from Tom's Mark IV).  The following are 
> histogram counts of
> two dark frames.  From h4r1836.524, I'm sure I can call 
> everything in the
> last bin as bad (they are all 65535), and for h3r1836.524, 
> the first and
> last bin are bad (0 and 65535), but what about "31207 to 
> 34328 =    2036",
> or any others?  All those 2036 values are the same (32321 in 
> this case, but
> they change value from each dark).
>