It's not easy being green - Image processing help request

[Entropy]

Good Morning everyone,

Due to being away from home, family and telescope for a long period of time I have decided to really try and put some effort into understanding PixInsight image processing. Also, I want to further practice using RickS' script he showed me at Astrofest last year (thanks again).
I have some data with me of the Eagle Nebula and surrounds in Narrowband I am trying to process but I am running in to some problems.
The Image has been successfully stacked and then I did a ChannelCombination using the standard Hubble Palette
after that i did a DBE then BackgroundNeutralisation then a ColorCalibration.
The image I am left with was very green. So I thought, no problem, and applied SCNR to the image. This left me with an image that appeared to be almost monochrome again.

My question is should I be applying SCNR to the Linear image then taking it non-linear and trying to bring the colour back in to it using RickS script or curves? or should I be taking the image to non-linear then trying to get rid of the green using RickS' script and then maybe applying SCNR if needed after I am finished? Or maybe something else entirely?

Attached are two images, before and after SCNR.
Thanks for your help

Scott

https://i.imgur.com/l2Fgcac.jpg

https://i.imgur.com/fFFUqh9.jpg

[Octane]

This is pretty normal because the area is so heavy in hydrogen alpha.

When it comes time to combine (I use PixelMath) I usually boost the other channels by a factor until the image starts to resemble something usable.

H

[Atmos]

When doing your colour balance in PixInsight for narrowband I find it best to turn for Structure Detect, this will stop it from trying to white balance on the stars.

I find that the white balance is best done once all of the stars are removed as the stars throw it off. As removing stars can be a painful process JUST to get a better balance, selecting a large nebulise area with very little stars or at least no big bright ones will give you a better balance.

[RickS]

Hi Scott,

A good way to balance the colours is to choose one as the reference, say Ha, and then apply LinearFit to the Oiii and Sii. Then do the colour combine and you'll have a much better balance as a starting point for further manipulation.

Also worth trying Colin's suggestion or H's. Lots of ways to skin this cat :)

Also FYI, SCNR has a much bigger effect on linear data than it does after a stretch.

Cheers,
Rick.

[multiweb]

:rolleyes: Don't listen to all those naysayer. Green rules. :evil:

[Entropy]

Thanks all for the help,

Now I just need to get through yet another 12hr shift at work before I can get home and give this a go.

And while I don't mind green, Marc... I think I may be over doing it on this one :D

[Octane]

Sorry, I forgot the LinearFit step as per Rick's suggestion, prior to messing about with PixelMath ratios. That's how I've been doing mine. :)

H

[sil]

Have you tried the new Photometric ColorCalibration? Which plate solves the image and uses accurate star colours to calibrate the image with?

[RickS]

The original question was related to NB images. PCC might produce interesting results but probably won't give accurate colour :)

[glend]

As was mentioned below, in the pallet you used green dominates because there is just so much of it out there, and you chose SHO as the pallet. The way you choose to display it in your narrowband images is entirely up to you, as the final image is an artificial construct. Boosting weaker channels pushes signal reality into abstract. Strongman Mike's recent Vela SNR is a good example of astro-expressionism, and is imho a wonderful image. There is no right or wrong in art, to the chagrin of some.

[sil]

Doh, sorry, Never had a chance to xperience NB for myself. Still shouldn't it work once youve combined them all anyway?:shrug:

[RickS]

The NB data isn't sufficient to reconstruct an accurate RGB image. I'll give it a go with some sample data and see what we get :)

[RickS]

Here are the results for a sample field (Eastern Veil.) They are RGB, Hubble palette (SHO) and Canada France Hawaii Telescope palette (HOS). It went the way I expected :)

[sil]

thanks :)