Author Topic: TGVDenoise Example: Noise Reduction in CIE Lab Mode  (Read 20757 times)

Offline Philippe B.

  • PixInsight Old Hand
  • ****
  • Posts: 399
    • CIEL AUSTRAL
Re: TGVDenoise Example: Noise Reduction in CIE Lab Mode
« Reply #15 on: 2013 May 15 06:37:49 »
What we need now is more data to test the tool. If you can upload an image, I'll be glad to make an example and post it here.

Thanks Juan, I will prepare you some linear monochrome images  8)

Offline gvanhau

  • PixInsight Old Hand
  • ****
  • Posts: 345
Re: TGVDenoise Example: Noise Reduction in CIE Lab Mode
« Reply #16 on: 2013 May 15 08:52:59 »
Hello Juan

Here you can download the image I used for testing:
http://rapidshare.com/files/1028361636/m83.fit

It is a linear image just after crop and DBE.

I also tested the tool after non linear stretch and it performed better.

Regards
Geert
« Last Edit: 2013 May 15 09:00:17 by gvanhau »
Geert Vanhauwaert

Offline IanL

  • PixInsight Addict
  • ***
  • Posts: 116
    • The Imaging Toolbox
Re: TGVDenoise Example: Noise Reduction in CIE Lab Mode
« Reply #17 on: 2013 May 15 15:16:17 »
I am starting to have some success with linear images.  The key is to work on the edge protection parameter first, and to use much smaller values than the default.  For example I am using 1e-5 for Chrominance and 9e-6 for Lightness.

I found the right values by creating a local support image first:

- Extract the Lightness and then use the histogram transformation process (with preview) to find the appropriate Highlight, Midtones and Shadow values.
- Cut and paste these in to the Local Support settings (quicker than repeatedly executing and undoing previews in the TGV tool to get the right stretch)
- Preview the noise reduction setting until you find a value that kills off any background noise in the protection image.

Next work on Lightness only (disable preview, disable chrominance):

- Apply an STF to the target image (tricky to find a good stretch for lightness only display when the image is linear but managed it in the end).
- Work on a small preview of the main image first, set iterations to 500 or 1000 and don't use convergence to start with.
- Start with default values and run TGV.
- You will get terrible ringing artefacts on a linear image so reduce the edge protection exponent by -1.
- Undo the TGV on the preview and reapply with the new exponent.
- Repeat the process of reducing the exponent, undoing and re-applying until you find a value for edge protection where you get no smoothing effect at all.
- Now perform a binary search between the lowest value where you got ringing and the value where there was no smoothing.  You should just need to work on the coefficient (or maybe go back up one exponent and then change the coefficient). 
- Start with a coefficient halfway between the lowest value you where you still got ringing and the no ringing values.
- If there is still ringing try the value halfway between the current one and the lowest value, conversely if there is no ringing go up halfway to the higher value.
- Repeat the binary search until you home in on the value where ringing just stops, which should give you maximum smoothing.

The range between ringing and no ringing is tiny in my linear images, the difference between 0.00000900 and 0.00000800 for example.

Then turn off lightness  process and follow the same process for chrominance.  It will be in a different range but probably very small again.  Once you have nailed the edge protection you can play with strength and smoothness to see what looks best.

Just be aware that once you switch to the full image using convergence you will get ringing again.  You just need to tweak the edge protection for lightness and then chrominance in the same way as before.  It takes longer on a full image but you should be a lot nearer to start with now.

In the sample below all the normal images are linear, starting with the stack which has had Canon Banding script, 2 x DBE iterations and initial ColourCalibration.  Next right is the result of TGV using values as shown in the dialog (plus slightly more agressive values for chrominance).  Next right is the extracted Lightness to the right as support and a clone of it showing the stretch applied by TGV.

Bottom left is an Atrous denoise on layers 3, 4 and 5 with a luminance mask as shown to the right of it.  The final result is a lot smoother than any other method I have tried on my noisy DSLR images and could be improved with a bit more tweaking on the Atrous settings I think.  Blinking between the denoised and original images, the features seem like they have been layered on top perfectly with no damage to the interesting stuff but good smoothing of the background, whereas most of the other techniques you can see a clear erosion of signal as well as noise to the outer regions of galaxies, etc.

Ready for the next steps in processing.

I achieved this in about 2.5 hours from a cold start after from reading the tutorial today, so this is an excellent new tool I'd say!
« Last Edit: 2013 May 16 04:47:51 by IanL »

Offline dgbarar

  • PixInsight Enthusiast
  • **
  • Posts: 91
Hi,

I have uploaded an image to endor for others to try their hand with the new TGVDenoise module.  Here is the link:

http://endor.uv.es/files/data/public/f31a3b212a820800c6ec7361030d4c2b.php?lang=en

Particulars about this image.  It was taken with an AP155 f/7 and an ST10XME at -5C.  This linear tricolor image of M101 was generated with 24 5 minute exposures through each of the RGB filters for a total exposure of 6 hours. Each of the sub exposures were calibrated with master bias and master flats generated from 50 subexposures.  The master flats were generated from 21 subexposures for each color.  All calibration, registration, and integration was done with Pixinsight 1.87.

The image is noisy. I have attempted using the new TGVDenoise module with this image in both linear and non-linear states.  As others have pointed it out, it is much easier to find TGVDenoise parameters when image is non-linear.  My best efforts with TGVDenoise still leave me with a "blotchy" background that I am unable to clip to my satisfaction.

It would be interesting to see what others can do with this image.

Cheers,

Don
« Last Edit: 2013 May 15 19:01:36 by dgbarar »
AP 155 EDF
Paramount ME
ST10 XME/CFW 8A

Offline Philippe B.

  • PixInsight Old Hand
  • ****
  • Posts: 399
    • CIEL AUSTRAL
Re: TGVDenoise Example: Noise Reduction in CIE Lab Mode
« Reply #19 on: 2013 May 16 00:26:30 »
Hi Juan

I sent you by email the link for these 6 linear images taken with my setup. THis is preprocessed integration.
I think they are typicaly what we obtain on deep sky nebulas objects with a CCD camera.



Most of them have very low light signal and moderate noise. Some have higher signal and noise.
The goal is to REDUCE (not REMOVE) the noise AND preserve maximum of the details.


TGVdenoise is a powerfull process but, as always, the job must be done during the acquisition (exposure time, number of exposure, tracking, preprocessing quality). I will not process a very noisy image.


thanks
Philippe

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 7111
    • http://pixinsight.com/
Re: TGVDenoise Example: Noise Reduction in CIE Lab Mode
« Reply #20 on: 2013 May 16 10:06:28 »
Thank you so much to all who have uploaded their images: Philippe, Geert, Don and Harry.

I have just completed a first example with one of Philippe's images, and will be working with the rest during the next days. These images will be of great help for us to understand and further develop the TGVDenoise tool.

Juan Conejero
PixInsight Development Team
http://pixinsight.com/