Hi Scott (@srosenfraz),
Thank you for your feedback ;-)
In the case of distortion correction, at the end of the registration, when the multiple view algorithm starts, a calculation is started that can’t be implemented using multiple cores. So you will only see 1 cpu core working, that’s correct. I don’t think I can make these algorithms multi-core. They are regression algorithms which will find the distortion parameters for all frames simultaneously. So 1 calculation calculates all parameters at once. This ensures that all panels are correctly related to each other. Especially when they have multiple overlaps with different frames. Maybe after more algebra courses, I can find a way to fasten them up ;-) I will also work on console output so you can better see that the algorithm is working and that the RMS error of the regression solution keeps improving.
The same holds for the LNC calculation. The normalization correction is calculated for all frames at once, so all frames are compared simultaneously in the calculation and the result will therefor be the set of correction parameters for all frames. Applying the correction to each frame is done using multiple cores already though. And the analysis of all frames before the LNC calculation starts is also done with multiple cores.
How is the mosaic working out? Let me know if you need assistance ;-)