Jump to content


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


whats your pi workflow for lrgb imaging

  • Please log in to reply
2 replies to this topic

#1 iwols


    Viking 1

  • -----
  • topic starter
  • Posts: 763
  • Joined: 27 Jul 2016
  • Loc: doncaster uk

Posted 04 December 2021 - 08:00 AM

well after watching hours and hours(and hours) of tutorials all which seem to do things slightly different ,just wondered what your workflow is, in as much detail as possible.My method is very hit and miss(mainly miss frown.gif ) thankswaytogo.gif

#2 imtl


    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 5,023
  • Joined: 07 Jun 2016
  • Loc: Down in a hole

Posted 04 December 2021 - 08:27 AM

It will depend on object as usual but there are some basic stuff. Maybe this will help you.

  • iwols likes this

#3 kathyastro



  • *****
  • Posts: 3,997
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 04 December 2021 - 08:29 AM

My workflow is mostly self-taught.  Most of the good information about PI is in videos, which I find annoying.  I hate giving 20 minutes of my life just to learn "What does this button do?"  However, I get decent results.  So I don't necessarily do things the "right" way.


I start by processing the Luminance.  The first step is a basic noise reduction using the NR feature of the Multiscale Linear Transform.  Then, depending on the image, possible Dynamic Background Extraction. 


Common wisdom has it that Deconvolution is best done in the linear state.  I have never gotten decent results doing that, but you do need to measure the point-spread function (PSF) for deconvolution in the linear state.  So I run DynamicPSF next and save the PSF image for later.


Then, I will use Histogram Transformation to stretch the image.  I seldom move the black point slider: there is too much data down at that end of the histogram.  I will usually bring the white point down until 0.001% of the pixels are clipped, which is only about 85 pixels with my images.  I bring the midpoint slider down until the histogram peak (which generally represents the background) is at 0.125 -  0.25 of the scale.


Next, I apply Deconvolution, using the PSF I saved earlier.


From then on, I work with the Curves Transformation and the Multiscale Linear Transform as required to bring out the details I want.  That includes lowering the background to about 0.125.  I would rather raise the background to 0.25 with the Histogram stretch and then reduce it with Curves so that I can clearly see and preserve the dim details.  I can get much more precise control with Curves.


Working with the RGB, I use more or less the same process.  I do the initial noise reduction with MLT on each master separately.  Then I will do a Linear Fit (I usually pick Red as my reference channel and apply it to Green and Blue) and a Channel Combination to produce an RGB image.  I carry on with DBE, stretching, etc., as for the Luminance.


Once I am happy with both L and RGB, I combine them using Channel Combination.  I select the L*a*b* colour space, specify the L image as the L source, and disable the a and b sources.  Then I drag the triangle from Channel Combination to the RGB image, to give me the LRGB image.


I will do whatever tweaking seems necessary, usually including a pass with ACDNR noise reduction, and then save the result.

  • iwols likes this

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Recent Topics

Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics