4 replies to this topic

[den1982]

Hello everyone,

 

 

is it possible to remove this gradient from the attached picture (RGb of R - G- B - Ha cannel combination)? I used continuum subtraction  to integrate the ha data into the RGB image.

I have tried GraXpert, DBE and Spectrophotometry-based Color Calibration in PixInsight. All led to even worse results.

Does anyone have any advice on which tool and settings to use?

Many thanks in advance

 

CS Dennis

Attached Thumbnails

• 

Edited by den1982, 03 December 2024 - 05:53 AM.

[vroobel]

It would be easier if you share the raw masterFiles. I assume all calibration frames were taken and used?

 

The ContinuumSubtraction script is not for the integration. 

Edited by vroobel, 03 December 2024 - 01:32 PM.

[Marcelofig]

PI's new MultiscaleGradientCorrection is imminent and looks quite promising. In the meantime you could try the current Gradient Correction and see what happens.

 

https://pixinsight.c...ion-tool.24528/

 

And it wouldn't hurt to post the image in the PI forum* and see if the PI developers can test the new MGC with it.

 

 

*In fact, the thing to do is to upload the image to an online service like Google Drive and post a public link there.

[den1982]

Thank you  for all the answers.

 

The new MultiscaleGradientCorrection tool sounds very interesting. I fond a solution for this problem:

 

My problem was, that I first used GraXpert to remove the gradient from my RGB image and after that I applied the color calibration tool. This results in that colorful background.
If I change the order and apply the color calibration first, everything is just fine.

[whwang]

I am afraid this will be a very challenging (or even impossible) case.  I see more than just gradient in your image, much more.  There must be something in your equipment that caused this.  The variation in background color has a size scale much smaller than the main target.  This will bring any traditional algorithms hard times on separating the background structure from the main target.  If I see such things in my image, I will immediately give up fixing it in post processing.  Instead, I will try to figure out why it is like that in the first place, try to fix it, and then re-image the target.

 

ps, It's quite possible that the new MultiscaleGradientCorrection tool in PI can fix this.  It is essentially using another good (or well corrected) image to replace the background in your image.  If the main background feature is a smooth gradient, you can say that the tool removes gradient for you.  If the main background feature is some bad imaging artifacts caused by your equipment (very probably like the case here), then it essentially replaces the bad part of your image with the good part from its database.  Purists can call this cheating.