WBPP with Separate CFA Channel Alignment and Noise Reduction

tpa

New member
This question is about how WBPP implements the Separate CFA Channel Alignment process. I am using WBPP v2.5.3.

When implementing the Separate CFA Channel Alignment process by selecting "Separate RGB" in the Post-Calibration tab, does the script automatically apply distortion correction and/or noise reduction or do these settings need to be configured under the "Image Registration" section under the Lights tab?

The original announcement regarding this technique (https://pixinsight.com/forum/index.php?threads/pixinsight-1-8-8-9-released.17106/) describes the use of this technique for distortion correction and provides excellent examples of the importance of setting the proper noise reduction parameters.

This post then states the following with respect to WBPP: "With version 2.2.0 we have the new separate CFA channel alignment feature directly available in the Weighted Batch Preprocessing script. To implement this feature the script applies a series of process parameters that we have found to work well in most practical cases, removing the need to perform the entire channel demosaicing and alignment procedure manually, as described above, except for difficult data sets."

This statement seems to imply that WBPP automatically applies both distortion correction and noise reduction with "a series of process parameters that we have found to work well in most practical cases" such that no further action is needed by the user to implement distortion correction with noise reduction. However, I wanted to confirm whether I understand this correctly or whether distortion correction and noise reduction need to be manually configured. I attempted to go through the logs to see what happened, but I was not savvy enough to figure it out.

Further, I note that selecting "Separate RGB" in the Post-Calibration tab does not update the associated settings in the dialog box under the "Image Registration" section under the Lights tab (distortion correction is unchecked and noise reduction is 0). If it's true that WBPP automatically applies "a series of process parameters that we have found to work well in most practical cases," will parameters entered manually be overridden?

Thanks!
Troy
 
Hello @tpa

I had the same questions and I did some digging. I believe this post is a little more explicit on the subject: https://pixinsight.com/forum/index.php?threads/wbpp-v2-2-released.17109/.
It states the following:

Once the "separate RGB Channels" or "Combined + Separated" option is selected, three grey masters will appear in the output list and be marked as "_R", "_G", "_B" in the Color Space column.
Files belonging to these three additional groups, which will be generated for each RGB group, will be aligned to the best reference frame adopting the following configuration:

Registration model = Thin Plate Splines
Distortion correction = enabled
Local distortion = disabled
Generate drizzle data = enabled
Noise reduction (Star Detection) = maximum value between two and the user-selected value
Compute PSF fits (Star Detection) = Distortion correction only

If I understand that correctly (and I did some tests which seems to confirm this) the settings "Distortion Correction" and "Local distortion" under the "Image Registration" section under the Lights tab are not considered if you select "Separate RGB" in the Post-Calibration tab. In this case "Distortion correction" will be enabled and "Local distortion" disabled regardless of what you set these settings.

It means that there is no way to enable "Local distortion" if you separate RGB channels. I wonder if this is normal? I have some cases of misaligned channels where I would like to try this option.
 
Back
Top