Hand Tracking Progress

We were able to get our hands on a Kinect v1 for the hand tracking portion of the project and it seemed to work pretty well!

That is, after an entire day (>8 hours!!) was spent installing libraries and getting both OpenCV and libfreenect working on the Pi.

Here's a demo of it working after all that time on a Raspberry Pi 3:
https://goo.gl/photos/fLYBDpTRrMvNqeDg8

The basic algorithm is as follows:

1. Get depth data from the Kinect depth sensor
2. Filter the depth data so only objects a certain distance from the camera are even considered
3. Find the contours of that data
4. Calculate the moment of each contour and from that you can calculate the centroid

All of the computer vision was done using OpenCV, which provides functions to do the various steps.