Producing flat frames

[fmw42]

Apologies. The value of the light frame pixels is reduced when dividing the lights by the auto-leveled master flat. I am then normalizing the lights to increase the pixel values.

I will try the other operators - so far median and auto-level produce flat frames that divide correctly and don't over or under compensate.

It is really hard to try to understand without examples. So perhaps I misunderstood.

But have you tried dividing by the median without the auto-level. Then after dividing, you might try using -auto-level, if it is still dark.

There must be a repository with the latest deb package.

Not likely. Linux distributions for most if not all flavors of Linux are way behind. You would probably need to install yourself from source.

[geoland]

That's no problem. Thanks again. Appreciate the assistance.

Still working on the master_flat/mode method which is a bit of challenge. In the meantime I need a work around that produces clean light frames with reasonable pixel values.

This works quite well - keep the processing linear. I found including -colorspace sRGB wouldn't divide cleanly, but noise estimates were better.

median combine the flats

Code: Select all

convert flats -evaluate-sequence median master_flat.tiff

light/master_flat and normalize the lights.

Code: Select all

convert lights master_flat -compose Divide_Src -composite -normalize

[geoland]

This writes the histogram to /tmp/mode.txt and cuts the mode value, as well as removing the leading whitespace to write mode as a variable $MODE

Code: Select all

convert *.tiff -evaluate-sequence median combined_flat.tiff >/dev/null 2>&1

	convert combined_flat.tiff -format "%c" histogram:info:  | sort -r -k1 | head -n 1 > /tmp/mode.txt
	
	mode=$(cut -d':' -f1 /tmp/mode.txt)	
	
	MODE=${mode//[[:space:]]/}

The next step is - pseudo code

Code: Select all

combined_flat.tiff/$MODE = master_flat.tiff

This does not work as well as hoped

Code: Select all

  convert combined_flat.tiff -evaluate divide $MODE master_flat.tiff

[snibgo]

From mode.txt, you are extracting into $MODE the number of pixels that have the most common colour. I don't know what significance this number has. Depending on your images, the number might be low (eg 2) or high (eg thousands or millions).

So dividing the image by $MODE will darken it, perhaps to almost or entirely black.

[geoland]

From what I read, the idea of dividing the flats by mode is to correct the multiplicative errors in the flat frames, before combining them. My application, above, is not correct because I am dividing the master flat (all the combined flats) by the mode. I realise now that the individual flats should be corrected prior to combination. Dividing the flats by mode should produce pixel values closer to 1. However, IM does a very good job of flat calibration of the light/image frames without complicating things with mode division. Which is not working.

I will keep working at this and report back, if things improve. Many thanks fwm42 and snibgo for your help.

[geoland]

I think this is solved. The order of the operators is important. This is an effective method of creating a master flat that does not over correct or under correct the light frame pixel values during division.

Code: Select all

convert flat.tiff -auto-level -evaluate-sequence median master_flat.tiff