4.2 Enhancements/Improvements for DSO Live Stacking

[admin]

Hi,

good to hear that the sizing issue is now resolved.

On the histogram side, does your image tend to have a hard black background (perfectly black or very close to it in many areas)? That would tend to tie in with the sort of histogram shape you are seeing in (presumably) the processed image with the histogram peaks very far to the left.

If you are using the new Arcsinh stretch - particularly with strong stretches appropriate for faint targets - it can be hard to get a satisfactory background without it being very black. The 'soften' adjustment for the arcsinh stretch is designed to help in this case - it makes the stretch weaker in the region that gives the bottom 20% or so of output pixel range, making it much easier to get a sensible background.

Thanks for the pointer on the CLAHE algorithm - I will have to look into that (probably after getting 4.2 released - that is getting quite close to complete now).

cheers,

Robin

[MarMax]

Robin,

In thinking more about CLAHE stretching and after more discussion on CN this may be something for the mini-histogram. Adjusting the histogram this way is typically for an image to be posted on the forum. For myself the preference is for a final live stacked image to have minimal color noise. To have your cake and eat it too some simple one-click option with the mini-histogram may be the best approach.

These "Save Exactly as Seen" files are 8-bit PNG files and include the mini-histogram adjustments. This could be a check box and include an associated autostretch. If you want to go back, just uncheck and reset or you can "Save with Adjustments" which does not include the mini-histogam adjustments and is an 8 or 16-bit PNG file depending on the camera.

MiniHistogram_wCHAHE.jpg (55.61 KiB) Viewed 92 times

And this is suggested as a future option. I understand you are focusing on the final 4.2 release at this point.

Mike

[MarMax]

It's too late to edit my previous post so I'm adding a bit more regarding my lack of understanding of CLAHE. It's sounding quite complicated and probably difficult to implement. And you have enough on your plate with the 4.2 release.

[admin]

Hi,

thanks for the further ideas - I have a note to look into this, and really I need to get a good idea myself of how it works, how it changes the image, etc before working out where it might be best to implement it. One of the image processing libraries I use does have functionality for CLAHE (but only for mono images), so I don't have to implement the algorithm to try it out...

Funnily enough, a few days back everything was quiet and I was thinking that the last 4.2 features were nearly done. This week though I have had lots of feedback with potential issues I need to dig into, so things are no longer quiet at all!

cheers,

Robin

[Jean-Francois]

Hello,

Concerning the CLAHE image processing ... it is now well known from the SHG users.
Mainly all the sun images processing is with a CLAHE contrast enhancement.

For the "histogram" processing ... you can have one "global" or one "local" processing.
The SharpCap histogram tools are all "global".

It exist some alternative or modified CLAHE algorithm, for example:
https://dev.to/muhammed_shahid_d7f50e64 ... -image-peo
https://arxiv.org/html/2604.16010v1
https://www.nature.com/articles/s41598-023-49534-y
https://www.sciencedirect.com/science/a ... 1026003171

Regards,
Jean-Francois