Dealing with large dark anomaly

[ecuador]

I shot the Virgo Cluster with a 135mm lens a month ago and I had forgotten to process it, but when I tried yesterday I could not find the appropriate flats for that session. So I have 5 large "dark anomalies", which I can't figure out how to deal with, as they screw up Wipe (and I really need wipe, they were shot from the city). What I tried is to put dark anomaly at max (30) and after a very long time, it still treated the anomalies like they are interesting objects with a nice nebula (light pollution actually) around them. I tried 50% binning (although I'd rather try avoiding it for this wide field image), and max dark anomaly correctly figured out 4/5 and missed the largest. Then I tried masking out the anomalies for wipe. That did not work either. So, any idea if there is something I can do besides even more binning to get all of them under the dark anomaly max size?
The file for reference: Virgo Cluster

[admin]

Hi,

If you know where the dark anomalies are, you can simply mask them out using, for example, the Lassoo tool. This way, Wipe won't sample them. Easy huh?

Screenshot-Untitled Window-3.jpg (101.66 KiB) Viewed 7435 times

Had a fair bit of trouble getting anything useful from the stack though...

[ecuador]

I had tried that as I noted, it was one of my first thoughts, but it turns out the problem was that I had made the mask too tight around the anomalies. I saw you were pretty loose about it and I tried it that way with success! I was also underwhelmed about what 25x2min @ f4 gave me, not sure what the problem is. Also after wipe there is a weird noise pattern in concentric circles if you noticed. I have seen banding with my DSLR, but that was new! I had tried the Orion Imaging filter on the camera lens, perhaps it makes things worse instead of improving when used that way...

[admin]

I had tried that as I noted, it was one of my first thoughts, but it turns out the problem was that I had made the mask too tight around the anomalies. I saw you were pretty loose about it and I tried it that way with success!

Cool. Yes, dust specks can cause a fair bit of havoc in the immediate vicinity. It's usually a good idea to be liberal with the masked out area.

I was also underwhelmed about what 25x2min @ f4 gave me, not sure what the problem is. Also after wipe there is a weird noise pattern in concentric circles if you noticed.

Yes, I did notice. This is usually caused by low bit-depth of the remaining data/dynamic range (e.g. the data/dynamic range that remains after removing the bias that Wipe detects).

I have seen banding with my DSLR, but that was new! I had tried the Orion Imaging filter on the camera lens, perhaps it makes things worse instead of improving when used that way...

If you figure out why the data rather poor for 25x2min let me know. I too am a little surprised. What camera is this?

EDIT: Actually, did you stack with median stacking (will retain DSLR's bit-depth) by any chance? If so, try stacking by averaging (this should increase bit-depth).

[ecuador]

If you figure out why the data rather poor for 25x2min let me know. I too am a little surprised. What camera is this?

This was a Canon 450D full-spectrum mod, with a Jupiter 135mm lens @ f4 with a 2" Orion Imaging filter attached.

EDIT: Actually, did you stack with median stacking (will retain DSLR's bit-depth) by any chance? If so, try stacking by averaging (this should increase bit-depth).

I am pretty sure I used Average. Actually I can produce an even larger stack, it is just that after the first 25 frames, a power cable entered the frame. I should be able to compensate by doing a kappa-sigma stack on DSS, however I am not sure how that would work with ST - DSS tells you to do a Per Channel or RGB bkg calibration with Kappa Sigma, but ST would not like that, right?

Thanks!

[ecuador]

By the way, perhaps some sort of "healing brush" for such anomalies might be a useful feature. There isn't something like that and I haven't noticed, is there?

[admin]

By the way, perhaps some sort of "healing brush" for such anomalies might be a useful feature. There isn't something like that and I haven't noticed, is there?

There is a definitely a method you can use here using the Heal module.

A healing brush is a very blunt/crude tool, and the Heal module is preferable for a number of reasons;

1. By putting all 'off limits' pixels in a mask, we can be sure no 'bad' pixels are taken as a source.
2. By putting limits on the replacement pixels we can be sure no new stars are generated.
3. By using the Layer module,we can restore any stars that were present.

The result is guess closer to 'reality' (though still ultimately a guess).

[ecuador]

Ah, ok, heal looks nice, thanks! Should I apply it as a first step, before wipe you think?

Also, any comment on what to do with kappa sigma dss? Use it without calibration despite the DSS instructions, or use rgb or per channel calibration and then tell ST that it is already stretched?

[admin]

Ah, ok, heal looks nice, thanks! Should I apply it as a first step, before wipe you think?

It's best to do it after you've processed the rest of the image. This way you can best assess the impact of the troublesome areas.

Also, any comment on what to do with kappa sigma dss? Use it without calibration despite the DSS instructions, or use rgb or per channel calibration and then tell ST that it is already stretched?

RGB per channel calibation does not stretch the image, however it does make noise levels no longer comparable between channels (which exacerbates noise in the aggregate luminance data).

I'm not entirely sure why DSS would want RGB channel calibration done...