Whoa, nothing so sophisticated, just commercial graphics software. Pasted the infrared image over the UV image, rotated it into roughly sorta the right orientation. Then I made it semitransparent so I could see how the top image registered over the bottom image, and moved the top image so the daisies overlapped--that was the central reference point. I switched to skew mode and just grabbed the corners of the image and pushed and pulled and tugged and generally skewed around until it looked right. Completely unmethodical and unscientific.

I'm sure that JPL has some fancy software that uses the pointing parameters of the two cameras to do a ton of trigonometry to move pixels around until they get a rough fit (but not as rough as my fit). That still wouldn't give you a perfect image, since some pixels will be hidden between images, so I still think they had to rely on merging the color information they got that way with one of the mono images to provide the brightness channel. It works because color has less spatial resolution than brightness, so the eye doesn't notice how the colors don't perfectly match.

Stereo viewing of separate color images--the stereopticon method--shouldn't be a problem. It is possible to create color anaglyphs, though the quality's pretty poor. You use red and green from one eye, and blue from the other.

Spherical distortion of PanCam images is very slight, and hardly worth correcting--PanCam panoramas fit together with almost no skewing. NavCam images, OTOH, have a wider FOV and exhibit a distinct "fisheye" effect. They're a bear to stitch into panoramas because of that. Doug Ellison's used some special software and, I think, gotten good results with NavCam images. You might want to ask him for the details.

Hope I covered all five or six questions...;-)