Data is collected in my backyard over the course of the winter, as I added more and more data, I have tested many different methods for processing, and this image is the "current" result, which I believe to be my best of my many attempts. 
https://www.astrobin.com/z6t0k6/K/

The image is an SHO image with RGB stars. 

The following process is used for all NB channels to prep each channel (SHO).
  • WBPP
  • NormalizeScaleGradient with Drizzle
  • Imageintegration
  • DrizzleIntegration - Kernelfunction = Gaussian, DropShrink=0.8, KernelGridSize = 20
  • GradientCorrection
  • DynamicCrop
  • BlurXTerminator (Correct Only)
  • LinearFit
  • BlurXTerminator (Sharpen non stellar = 50%, sharpen stars = 10%)
  • StarXTerminator (Large Overlap)
  • NoiseXTerminator (Denoise=40, Detail=15)
  • Manual stretch with HistogramTransformation
  • Finetune with CurvesTransformation
  • Finetune details with GHS


Sii stretched - https://www.astrobin.com/z6t0k6/L/
Ha stretched - https://www.astrobin.com/z6t0k6/M/
Oiii stretched - https://www.astrobin.com/z6t0k6/N/

I combined my NB channels with ChannelCombination and extracted the CIE L component.
The L has been only slightly modified:
  • NoiseXTerminator
  • CurvesTransformation


L Component - https://www.astrobin.com/z6t0k6/O/

I created an ACDNR mask of the light image to target all the background I would later need to control.
I discarded the RGB image again.

I then recombined my NB channels with ChannelCombination into a new RGB image.
The RGB image has been processed in the following way:
  • Bill Blanshans, NarrowbandNormalization (Preserve Lightness, SHO Palette)
  • CorrectMagentaStars to remove Magenta residual from the image
  • Created a variety of color masks (Red, Blue, Yellow, and others that targeted very particular areas of the image in various Red and Yellow colour strengths)
  • Curvestransformation with the different color masks, to accentuate colors and areas in the image.
  • LRGB combination with the L component
  • CurvesTransformation with previous inverted ACDNR mask to "dial down" background light a bit.
  • Local Histogram Equalization (Radius 240, 12 bit, slope limit 2, amount 0.150)


By using color masks I find that I have more control over how the colors are controlled in the image. 
I have tried Bill Blanshans scripts, NBColorMapper and others. 
Above is heavily inspired by these 2 Youtube videos: 
https://www.youtube.com/watch?v=cn7UGWlPxfI
https://www.youtube.com/watch?v=fo6E0pBvdLQ

The RGB Image was now loaded into ImageBlend:
  • Created a High Pass filter image
  • Blend High Pass Image to original to sharpen contrast
  • Created a Darken/Min and Lighten/Max image from newly created sharpened image
  • Blend Darkened and Lightened image to a Final blend.


Apply RGB stars to final blended image using ScreenStars Script. 
Reduced stars with Star Reduction script. 

That's the used workflow. 

Input etc. are more than welcome.
Like