Fixed pattern noise?

[pfile]

upload to google drive, dropbox, etc., allow the file to be read by anyone with the link, and post a link here.

rob

[silentrunning]

Ok

Hopefully you can now download it...

https://drive.google.com/file/d/1USCgBPJQNNKYIqiNQj-kdrGbrghV5ryF/view?usp=sharing

John

[pfile]

yes - the link works

rob

[vicent_peris]

Hi John,

I tried the script in your image and it does not work. The only reason I find is that the lines are not completely vertical. So, IMO, the fourier solution is the best one.


Best regards,
Vicent.

[silentrunning]

Hi Vicent

Thanks, this is all very educational - I guess the dithering and integration essentially blurred what are residual vertical lines left from the bias.

If I perform an FT on my master bias and delete all but the middle horizontal line from it  and then perform an inverse FT I get what appears to be a model of the vertical data in the bias. I suppose if I could somehow align N copies of this vertical line model by using registration data from the lights I might end up with the same blurred vertical bands that appear in the integrated image - which came from N frames - which could then be subtracted out.

It's just out of interest that I'm pursuing this now but is there a way of acquiring the registration information from one set of data and applying it to a second set?

Many thanks

John