Questions on "LocalNormalization" process

[Philippe B.]

Hello


I have question on LocalNormalization process


1) On which image it mus be done ? is it on calibrated cosmetized images , or on registred ones ? I see there is a generated file which process like drizzle which will process from calibrated ones. In fact, I mean I would go for registred image ones (this is more logic I think) but I have doubt if there is black bands after alignment which can affect large structure effect  


2) If I have S2, Ha, OIII, R, V, B images, should I choose a reference image for each stack (ie : the best Ha for Ha stack, the best OIII for OII stack, like I am doing for ImageIntergration of each stack. (of course StarAlignement has a common reference image (the best of all)
 
thank you !!!


cheers

[mmirot]

I am not 100% sure on this.

My tests were done using the best calibrated image as a reference.  It has to be the same filter.
If you are integrating CC frames, I would use a CC  & Calibrated frame as a reference. This just makes sense to me.

Max

[mmirot]

I forgot, it should be registered to the stack too. What ever pre-integration steps you normally use should be followed.
I don't use dizzle so I don't know what to I tell you.
 My guess is use a registered and drizzled frame as the reference as this is a pre-integration step.

Max

[Philippe B.]

After several tests
- must be done on registered images (if not, error message)
- should be done with a reference of each stack


difficult to find the right scale. For example 256 give black holes and 128 got a blotched background. 192 seems a compromise.
Better result in Windsorized than Linear Fit. It is a stack of 100 images.

[Juan Conejero]

Please bear with me. I'll answer your questions later; now I have other commitments and pending work that must come first.

[Philippe B.]

Don't worry Juan... 
As always when there are some new features in PI, I want to test... I check all parameters with different values (default, lower, higher, then adjusted ones... ) and see what happens ! Even nice or bad results, it permit to learn how to handle this process.


I will wait for your tutorial, for sure,...