I like the pneumatic piston idea. I was thinking of a frame with multiple pneumatic controlled rods. The frame would travel in steps and be mounted on a larger frame that connects to the limb. The smaller frame assembly would hold the pneumatic detector rods. It would scan one horizontal.. line on each pass. Total lines needed would be determined by travel distance and step increment.
The pneumatic detectors would find the zero position of the rods with a small amount of air pressure. A computer would read each rods zero position. The 'read' pressure would then be applied and travel distance noted.
The system could easily display the 3d surface of the limb for zero and read pressures. I would also suggest rotating the whole system and doing multiple passes.
The position detectors would be very simple and accurately detect any position of the rods. Imagine inserting a long steel core being inserted into a long electromagnet. This could change either impedence or frequency of an oscillator. Positions could easily be logged and linearized.
Rick