Hi Dennis,
I had the same thoughts about capturing PTMs and RTIs with plenoptic cameras when I heard about the Lytro, and I had a little discussion with folks at CHI about it. The proprietary software in the Lytro limits its application, unfortunately.
Marc Levoy at Stanford, with students and collaborators, is developing a Linux-based, open architecture, programmable camera for computational photography they call the "Frankencamera." It's not pretty, but it appears it could possibly be developed into a plenoptic camera that could be programmed to take PTMs and RTIs. My understanding is that plenoptic sensors are currently limited in resolution, but I'm sure they will improve. More information about the camera is here:
https://graphics.stanford.edu/projects/camera-2.0/Best,
Taylor Bennett