Kinect Chapter 12.   Motion Detection Using OpenCV

This chapter was previously labelled as "NUI Chapter 16.9".

[Note: this chapter is virtually identical to NUI Chapter 3, except that I'm using the Kinect sensor as the input device rather than a webcam.]

This chapter explains how to use the Kinect's camera to detect change or movement in a scene. Frames are grabbed from the camera and drawn rapidly onto a panel, using the code developed in the previous chapter. A movement detector analyzes consecutive frames and highlights any change/movement with a pair of crosshairs at the center-of-gravity of the motion. The application, called MotionDetector, is shown above.

The detection algorithm is implemented using the OpenCV computer vision library. OpenCV is such an important tool that I'll spend a little time explaining its capabilities, and especially its Java interface (called JavaCV), before describing the detection application.

Downloads

• The PDF file for the draft chapter (640 KB). Last updated: 18th December 2011.
  
• Zipped code (52 KB). Last updated: 18th December 2011.
  IMPORTANT 1: the installation details for the OpenNI and NITE software have changed since the book was published; please read the details on the main KOPS page.
  IMPORTANT 2: Please read the notes on which JavaCV software to use, located on the main VBI page.

• Up to the Kinect Open Source Programming Secrets Index Page
• Back to Chapter 11. Kinect Capture
• On to Chapter 13. FAAST-style Body Gestures