Re: [APML]: Computer Enhancement

["Tom Krajci" <krajcit@det5wg57.barksdale.af.mil>]

> From:          celstar@wi.frontiercomm.net

> . . .Personally, I feel that enhancement 
> should be used sparingly and true to the original image, not to add 
> color that isn't there, or remove star trials due to guiding error. 

Strictly speaking, in the original 'trailed' image due to bad 
guiding. . .most of the information is still there as to what an 
untrailed image looks like. . .and FFT processing techniques 
(processing in the frequency domain) just rearrange this data so 
that it's easier for us to view it in the spatial domain.  If you 
look at the math involved there's no "just make up some of this 
stuff!" function involved in FFT processing.  

Same with boosting the color saturation.  The color's already there, 
not being 'colorized' like some old black and white movie.

Color imaging is tough.  Calibrating films, printers, monitors, etc. 
is also tough.  Your eye is very adaptive to color.  Set up three 
color test charts side by side, but illuminate one with sunlight, one 
with flourescent light, one with tungsten (incandescent) light.  Your 
eye/brain will note subtle color differences that probably aren't 
annoying. . .but film/video/CCD this scene and all hell breaks loose!

Tom Krajci

Capt Tom Krajci
B-52 Intelligence Officer
"Trust. . .but verify!  (Soviet maxim)"
http://spur.barksdale.af.mil