Key in each of the four Processing examples at the end of the Vision chapter (p. 547) and experiment with them. Understand the basic idea behind:
Frame differencing Background subtraction Brightness thresholding Brightness tracking (a penlight helps nicely here)
Run all of these, but then focus on one algorithm to experimentally modify and write a short technical document describing what you tried and the results (remember, by their nature, experiments are allowed to fail). Post example screenshots to the wiki and be prepared to demonstrate in class.
I expect to be wowed by the creativity exhibited in your experiments!
For example, we can apply any image processing algorithm we like to each frame of the video stream. Can we apply such algorithms to just the detected motion in a stream?