12 2013

Facetracking with Kinect December 30th 2013 / Mayco

Recent experiments with OpenCV and the Microsoft Kinect made me interested in OpenCV (or “computer vision” in general). The Microsoft Kinect offers some unique features over typical web cams (at least at its price point and availability). The feature that interests me the most is the ability to get a “depth frame” from the Kinect. Using a depth frame it’s possible for the computer to “see” depth and color (from the RGB camera of the Kinect) as opposed to just color.

One of the things I wanted to do was combine computer vision with robotics. To get a proof-of-concept up and running I modified a Kinect face tracking example so it can communicate with an Arduino Uno hooked up to two servos in a pan/tilt configuration. When the software detects a face it converts the X/Y coordinate of the face to a format usable for the two servos and sends the information to the Arduino over a serial connection.

Eventually I would like to be able to track objects using a combination of both the color and the depth information instead of a face tracking. The face tracking of the Kinect is pretty good for what it is, but it’s still far from perfect. Tracking objects seems more versatile than just tracking faces, so that’s what I’ll be looking into next.

Back to projects