Now I'm using a USB microscope and physically changing the focal point with a stepper motor driving a large laser-cut gear. Next I'll hook an EasyDriver I have laying around to an Arduino and start capturing some test focal stacks. Using those I should be able to finish the software that converts the focal stacks to 3D models. Less immediate concern is to package it more nicely so that I can orient and position it precisely.
This scanner will be at the heart of a sorting machine: