Dave,
once I generated an RGB image, I had to extract the channels so that I had something to use with the LRGB function
No, you don't have to extract the individual RGB channels. Follow these basic steps:
- Open your RGB and L images. The RGB has been previously combined from the (registered) individual RGB frames (with ChannelCombination) if you work with a monochrome CCD. Or if you work with a DSLR or one-shot color CCD, you already have a RGB color image after calibration, deBayering, etc.
- If you haven't already registered RGB and L, do it now with StarAlignment.
- Stretch your L and RGB images with HistogramTransformation. Try to achieve approximately the same brightness and contrast for both images, but don't try to do a very precise work; we'll do it automatically much better in the next steps.
- Extract the CIE L* component of RGB with the ChannelExtraction tool (select the CIE L*a*b* space, uncheck a* and b*, and apply to RGB). You can also use the
Image > Extract > Lightness main menu shortcut.
- Open the LinearFit tool (ColorCalibration category) and select the L* image you've extracted in the previous step as the
reference image. Apply the LinearFit process to your L image. After this operation, you can close the L* extracted from RGB since it is no longer needed.
- Open LRGBCombination. Disable (uncheck) the R, G and B channel slots. Leave only the L slot enabled and select your L image clicking the down arrow button. Be sure the lightness transfer function parameter is at its default 0.5 value (or you would be destroying the perfect match achieved with LinearFit). For the same reason, be sure the four channel weights are set to 1 and the "Uniform RGB dynamic ranges" option is enabled. You can play with the saturation parameter if you wish (decrease to increase saturation). Apply this process to your RGB image, and you're done.
Now you have an optimally matched LRGB combined image. Note that the LRGBCombination process can be used with previews. So you can define one or more previews on the RGB image over areas of particular interest, and try with them to quickly find a result that you like; then apply to the whole image.
Let me know of it works this way.
