Some weird questions

[turboscrew]

Does anyone have good pointers in mind about the actual methods of digital imaging and enhancement?
I'd need to know at least about automatic gain control.
Gamma correction and non-uniformity correction would play an important role too.
Also contrast limited histogram equalization would be of interest.
I'm supposed to implement that in an embedded system.
Google haven't been my friend on this.

[SkyHiker]

Is this for astronomy or other use? For AP you have to contend with read noise, exposure time and gain per filter. Read noise is negligible these days, and stacking can help increase the dynamic range by choosing relatively short exposures with high gain (but low enough to not saturate the stars).

For daytime photography most modern cameras can take 3 pictures sequentially quickly at various exposure times and/or ISO that can be combined into one to increase the dynamic range.

A lot depends on the sensor and the noise, which are camera and environment specific. I know, that's not of much use. So here are some links for further reading. Craig Stark is excellent for AP, check the various articles at http://www.stark-labs.com/craig/articles/articles.html . I like ClarkVision too, https://clarkvision.com/articles/digita ... index.html . Here's a tool that automatically selects the best ISO and exposure time for AP, https://sourceforge.net/projects/astrophotographylab/ . Here's an interesting link for upstream and downstream noise, also with ISO recommendations: http://dslr-astrophotography.com/iso-ds ... otography/ .

I may look into this myself as I am looking for a topic to present at the Kaibab Lodge star party in June.

[turboscrew]

It's for other use, but I guess most of the stuff applies. The sensor is a thermal image sensor, and it's like EAA - from image sensor to eye in real time. There's a special HW for speeding up the pixel manipulations.
There will be no OS, so ready-made libraries and applications are out - except sources for reading about the algorithms.

The system is supposed to take darks frames every now and then, and do a bad pixel replacement.

Maybe I can use that knowledge later in some hobby project for astronomy.

[SkyHiker]

The last link links to this paper where the math is summarized in easy reading: https://homes.psd.uchicago.edu/~ejmarti ... ead_vs_iso . That may be the easiest lead to something implementable.

[turboscrew]

The last link links to this paper where the math is summarized in easy reading: https://homes.psd.uchicago.edu/~ejmarti ... ead_vs_iso . That may be the easiest lead to something implementable.

The PRNU in the paper is probably the same as NUC (that I need to implement).
The NUC and AGC are, probably, the toughest parts for me,

[turboscrew]

Just for clarification, the sensor here is really just the sensor, not a camera.