Hi.
Sorry for the late reply - just stumbled across this. I am currently experimenting with this, and may have something useful to you.
I have several batches of 10 images of a 6 segment near full moon. For each segment, if I use the script/utilities/fftregistration script, it will register the lunar segments reasonably well, and output relevant .fit files.
I am experimenting with using the standard ImageIntegration routines to combine these. The results are almost promising - I do get an integrated image output, but with 10 input images (per segment) the algorithms don't seem to be coping with seeing differences between frames very well at all.
I am hopeful that there will be a selection of options that are useful, but somehow not holding out much hope...
PS: I'm using PI 1.8-RC5 for this. Previous attempts with fftregistration script (circa PI-1.6 or so) weren't this successful.
Cheers -