Cause of blotchy color background noise

[mborland]

I followed the tutorial at https://www.lightvortexastronomy.co...rating-and-stacking-images-in-pixinsight.html to stack 63, 180s subs of M101. After all is done, I get a blotchy color background as shown in the attachment. I didn't succeed in using noise reduction techniques (e.g., multiscale linear noise reduction) to get rid of this: it just turns into a diffuse elevated background, mostly.

I'm unsure of the source of this noise or what can be done about it. Is it just light pollultion and moonlight? Any advice would be appreciated.

Imaging conditions: Bortle 8 (Chicago suburbs), full moon, Optolong L-Pro filter, 8" Meade LX90, ASI071MC Pro at -25 C, CEM40EC mount, off-axis guider with ASI290MM mini. As stated, the subs were 180s each and I stacked 63 of them.

Thanks--Michael

 

[szymon]

That looks very much like my light pollution from an OSC camera in horrible Bortle 8 (mine is London not Chicago). It's a lot of work to process out but it is possible. I look forward to seeing what advice people give you

 

[pfile]

man i just spent like 10 minutes searching for it, and i can't find it, but i believe Alejandro Tombolini had a tutorial where he masked off the foreground of a galaxy image and then used the Atrous wavelets tool in Chrominance mode to remove the color blotches from the background at higher scales.

if you look at http://pixinsight.com.ar/en/ he has quite a number of tutorials where he works on background chrominance noise, but because the tutorials are somewhat old he's using ACDNR everywhere. you can probably use the same masking techniques but then use TGVNoise in chrominance mode instead, as TGVNoise is considered more advanced than ACDNR.

but yea that's typical of LP imaging, my images look like that too.

 

[szymon]

Yes, this is exactly what I did, TGVNoise is what I used (following instructions from someone much more experienced than I!).

 

[mborland]

Thanks, all. I'll do more searching and experimentation with different noise reduction techniques. At least I'm not alone in confronting this problem.

--Michael

 

[dld]

Hello,
I wonder if you have used LocalNormalization. If yes, try integrating without using LocalNormalization.

 

[mborland]

Hello,
I wonder if you have used LocalNormalization. If yes, try integrating without using LocalNormalization.

Thanks for the suggestion. It certainly changed things, but not clearly for the better. Instead of somewhat random blotches, I see ripples, as if the galaxy just plunged into a pool of water.

 

[ngc1535]

Are you *really* undersampled? I calculate you are imaging at 0.5"per pixel. To be undersampled at this platescale is pretty hard. If you were 2-3" per pixel I would think it would make sense. Instead of Drizzling- you may find that simply upsampling your image provides the same image scale you are looking for. I would recommend you do the experiment and see if Drizzle is actually giving you anything beneficial.

The reason I mention this... if you do not Drizzle... you can add another tool to your noise arsenal- MURE Denoise. You would break your color image into the "half-sized" Red, Green, and Blue mono-channels... then you can apply MURE Denoise to each. Then you can combine to create your color image..and upsample from there. It might be the noise reduction at that point..plus TGVDenoise later will be beneficial! It is a different route to consider if you are adventurous.

Even with the image in the state you have it in- you can preferentially use TGVDenoise in Lab and go crazy on the color side- which will smooth and desaturate things. Of course simply attacking the background with desaturation can probably do some good here as well. You just need to protect the good stuff.

Chicago plus the full moon is the real problem. You just need more signal/darker sky... of course the Chicago part is tough.

Just some things to think about!
-adam