Bad Pixel Map?

[Hans]

Hi all,

Jasem Mutlaq wrote on 20210313:
> We have "dark" subtraction available in KStars for the align/focus/guide
> modules. I was about to begin implementing a new way to build a dark
> library for images. However, Han pointed out that using bad pixels maps
> actually may yield better results, and it's a lot simpler than taking all
> the darks and stacking them using different combinations of
> binning/exposures/...etc. Therefore, I put the dark library revamp on hold
> as I investigate this venue.
> 
> PHD2 uses both methods to remove noise from images. I believe it would
> substantially improve focus/guide/align modules if we can introduce a
> SIMPLE method to remove MOST of the noise.

I found the bad pixel map in PHD2 to be vastly more
effective than a master dark, and faster to make too. And
easy to update when a new bad pixel arrives over time.

It's quite well explained here :
https://openphdguiding.org/man-dev/Darks_BadPixel_Maps.htm

> Remember this is used to help
> these modules operate correctly, it's not used for post-processing images.
> 
> What do you think? Shall we adopt bad pixel maps to remove hot/cold pixels?

The idea is that bad pixel maps work better for uncooled
guider camera's.

> Do we need to do this like darks for multiple exposures/temperature/...etc

And a bad pixel map works for all exposure times and 'the
usual' temperature range.

> and save the results somewhere? or a "master" map would be sufficient to
> get the job done?

Yes, a master map. See above URL, it is made there by just
10 'long enough' exposures. With 'long' being longer than
the usual guide exposures, PHD2 lists > 10 sec for that.

> Hy previously introduced "Show Clipping" to highlight saturated pixels in
> the image, could this be used as a startup point? I presume something akin
> to PHD2 Bad Pixel Map (BPM) wizard?

Also add the cold pixels.

-- Hans