repairing reverse blooming lines

[wimvb]

(A short presentation of a processing step that may come in handy.)

My camera has all kinds of features to teach me image processing: high noise, hot pixels, amp glow, and on bright stars even reverse blooming.

Reverse blooming is when dark lines emanate from bright, saturated stars. Normal blooming causes the light to spill over in other pixels of a (CCD) sensor row or column, while reverse blooming causes dark lines. I have seen it in several instances, and it seems most common in dslr's:

http://www.cloudynights.com/topic/518196-troubleshooting-image-artifacts/?hl=blooming#entry6886422

http://www.cloudynights.com/topic/359253-problem-with-1100d/page-2?hl=blooming#entry4656068

Here's the effect on one of my latest image.

Dark lines emanate from the brightest stars in M45 towards the left of the stacked and stretched image.

Here's a crop of the offending section.

Blemishes like these will ruin any chance of pulling out faint detail from the background. Unless ...

Alejandro Tombolini (pixinsight.com.ar) showed in one of his processing examples how he removed one such line in an image.

Since the lines only cover a smaller part of the image width, CanonBandingReduction doesn't work. It will only introduce other faults. I adapted Tombolinis method so that it also works on lines that do not run the entire image width.

The short version is that I create previews to which I apply CanonBandingReduction. I then blend the previews with the original image. The longer version and how to of the method is in my blog.

It takes a while to isolate all the lines and find the best settings for the banding reduction, but I think that I managed to do quite well.

Here's the processed image.

There are still weak traces left after processing, but they don't interfere with the main target.

I hope you will never need it. But if you do, now you know at least how to remedy it.

Cheers,