Impact of Deconvolution on LRGB Images

[Alejandro Tombolini]

Hi Don,

Hi,
I did a search on endor to determine how to make a request for an account.  Did not find anything accept this thread.  How do I make this request to Juan?

See this:

To use PixInsight Files you must be a registered user of this new service. If you want to register, please send us a request at any of our support email addresses, or use the contact form on our website. To help us identify you, either use the same email address that you have associated with your license, and/or specify your user id. You can find your user id (a 16-digit number) in the email you received when you purchased your license, and also on the Resources > About PixInsight dialog of the PixInsight Core application. Unless you tell us otherwise, your PixInsight Files account will have the same user name and password as your current PixInsight user account (which you use to access our Software Distribution interface).

BTW.  I am starting to make some progress on this issue by going the Parametric PSF approach.  Not quite there yet.  I know what the parameters mean.  Just not certain how to optimize them to get the desired result.

I suggest that you choose the best stars in DynamicPSF, not those in the corners.

Saludos, Alejandro.

[dgbarar]

Thanks to those that have been helping me with this problem.  I think I have another piece to the puzzle.

I have been playing with parameters in Parametric PSF in Deconvolution.  The base parameters are Standard Deviation = 2, Shape = 2 (Gaussian), and Aspect Ration = 1.  If I adjust the standard deviation to 1.5, I get a halo around the star in question that looks very similar to if I use the exported PSF generated with Auto PSF Model.  My guess is that with PSF Model Functions being set to Auto, the exported PSF is under predicting the standard deviation.

So my thinking is that I need to adjust the PSF Model Functions for the exported PSF.  Does anyone have thoughts on what might be a good starting point.

Don

[dgbarar]

Hi Alejandro,

Thanks for the information on how to upload a file.  I will work on that.

With regard to star selection for the PSF, I followed the guidelines in Juan's thread from July 2011 introducing the New Dynamic PSF tool--you had suggested this thread earlier.  I redid the PSF just to make sure, but this time I used brighter stars A = 0.10 - 0.32.  I have attached the .txt file for yourself and others to review.  The synthetic PSF that was generated and subsequently used in the deconvolution still resulted in halos around stars.

Don

[dgbarar]

Hi All,

I have uploaded three FIT files to endor.  They are named M15 Blue, M15 Green 180, and M15 Red 180.  The trio of stars we have been discussing can be found on the Red at (924,456), Green at (924, 453), and Blue at  (900, 456).  Since this the first time I have uploaded files to endor I have absolutely no idea on how you find them.  I am guessing by User ID.  My User ID is dbarar.

Let me know what you think.

Cheers,

Don

[Alejandro Tombolini]

Hi Don, I cannot find the files. Make sure that they are in "Forum Shared Files" and not in "My Files".

Saludos, Alejandro.

[dgbarar]

Hi Alejandro,

OK.  They should now be in forum shared files.

Don

[Alejandro Tombolini]

Hi Don, I found the files and could download them without problem.

As I usually work with RGB images, first I align the images taking blue channel as reference



RGBCombination



Then a crop to eliminate borders



DBE to eliminate gradients



With AberrationSpotter make an image of the corners to be use as reference image in BackgroundNeutralization



Background Neutralization



ColorCalibration



Select stars to determinate the PSF



Starmask to protect cores of stars during Deconvolution



Image previous to deconvolution. Starmask has been activated to protect stars cores



Deconvolution selecting the PSF with values by default and I changed the algorithm to Regularized Van Cittert because I like the sharpen that it gives, particulary in globular clusters.



Now increase number of iteration to 20



To solve rings activate deringing and try with values by default



Better, but it needs even more global dark, I continue trying until find the value that let no rings.



Now watching in another area can be seen pixels in the background that this algorithm highlights but are not stars. Examples in preview02



Then in Wavelet Regularization I increase the Noise threshold in layers 1, 2 and 3, by try and error until find the values than correct it



Comparison



And the process may continue with the stretching. I would apply Masked strecht, may be in to parts with HDRWavelets in between.

Now, as I suspect that you will not be completely satisfied with the result, because it is noticeable a blue halo in the big star in the previous image, I will try and easiest way.

Apply ATWT to the blue channel and correct the halos before RGBCombination.



Then repeat the same process than before. See the big star near the center than had the blue halo and now it disappeared.



Saludos, Alejandro.

[dgbarar]

Hi Alejandro,

First, I want to thank you for all of the time that you have spent to help me improve my image processing.

I see a number of processing strategies that you have employed that are different than what I have been using.  These include:

- Deconvolution on the RGB.  I had been deconvolving the individual R, G, B channels and then compositing to an RGB.
- ATWS sharpening on the blue channel to reduce star size.  I had only been sharpening the entire RGB.

I need to study your process methods more and will report back in a day.

Warmest regards,

Don

[wnourse]

This thread is a perfect example of why I love the PixInsight community!  The amount of time and effort that people put into creating detailed explanations of their processes has helped me immensely in learning both the theory and practice of image processing with PixInsight.

Thank you!

[chris.bailey]

that trick of doing ADWT on a single channel to reduce halos is a winner indeed!

Chris

[astrodoc71]

Absolutely impeccable timing on this thread as I am just going through processing M13 with related questions! Thank you all!
Regards,
Dave