I first started thinking about the "chained brightness adjustment" technique back at the time of Pathfinder. The usual practice (and I'll bet NASA uses it with the BonnePan) is to average the brightness of all images, which will discard the useful information about hue/saturation as a consequence of sun azimuth. I finally decided to try it out with a quick-and-dirty and dog slow Visual Basic program ("dog slow" and "Visual Basic" are redundant when you talk about image processing). I had to run the entire procedure several times to fine-tune it, and each run took about half an hour. So, all in all, it was worth the 18 hours to finally get to try it out.

Looks like that PT software does a nice job stitching images and making geometric corrections. PanCam images are easy to do by hand because the optics are relatively narrow, and thus don't have much fisheye distortion...horizontally. Your ambitious plan for a 3-row panorama is going to require a lot of skewing on the bottom row.

The wide-angle NavCam images are impossible to stitch by hand, and PT will no doubt shine with them. NASA's NavCam panoramae have this looping or scalloped border where they had to do spherical correction. You could carve out quite a niche applying PT to NavCam.

Holger Isenberg in Germany has set up a site to automatically generate color from PanCam, at http://mars.gh.wh.uni-dortmund.de/mer/. I've talked to Holger about it, and he's using some simple recipes based on the available filter sets. The BonnePan images are near-infrared (L2), green, and blue, which gets you into the ballpark. Nick's baseline recipe of a 42% reduction in saturation seems to work well, although as I mentioned saturation varies from huge at the ends to minor in the middle. You'll have to play with it.

I'm planning to organize a gallery section on rovers.habitablezone.com. Looks like you already have Web space handled (image-by-image), but I'll be sure and put in a link if you can give me a central index page URL. You're doing nice work.