Quote:
Originally Posted by stanlite
Any comments welcome particularly if anyone knows how to use deconvolution in PixInSight lol
|
Hi Grady,
I recently started using deconvolution in PI and I'm pretty happy with the results I'm getting. I read a few posts on the PI forum about deconvolution and then played around and developed a simple workflow.
Here are a couple of the posts I found helpful:
http://pixinsight.com/forum/index.php?topic=4048.0
http://pixinsight.com/forum/index.php?topic=3261.15
A quick summary of what I do:
- Using DynamicPSF I select 70 or 80 stars from across the FOV that don't look too large or too small then winnow out any that are too saturated or faint (sort by Amplitude), vary too much by MAD, have a Theta or Aspect ratio that looks out of whack. Then I generate a synthetic PSF.
- I build a star mask that will protect the largest stars as a deringing support. Usually needs some tweaking of midtone parameter since the image should be linear at this stage.
- I build a mask that will protect areas of low signal from the deconvolution by cloning the image and using HT stretch with some clipping and maybe a touch of Curves. Apply inverted to the image.
- Set parameters for the deconvolution (set the PSF, deringing support, tweak the global dark parameter - usually between 0.01 and 0.03)
- Test against a preview with different numbers of iterations, tweak global dark, etc.
- When happy run the deconv over the full image
It sounds like a lot of work but it's not too bad with a bit of practice.
Cheers,
Rick.
