OpenCVKinectShapeTracker

What does this do do?

This example shows how to detect sphere like objects from the Kinect using a series of OpenCV steps

1. Open a Kinect 2 color and depth stream
2. Map the color information onto the depth information
3. Segment the color frame using EGS image segmentation algorithm and ofxImageSegmentation
4. Run contour tracking to determine geometric properties of each segment
5. Filter out small, oversides, and oddly shaped segments
6. Sense depth positions of the valid segments to find their 3D positions in space

How do I get it to work?

NOTE: this example is currently Windows only becaue of Kinect Common Bridge dependency.

1. Download openFrameworks 0.8.4 for VS2012
2. Clone this repository into apps/myApps
3. Clone the following addons into addons/
4. ofxKinectV2
5. ofxUI
6. ofxCv
7. ofxImageSegmentation
8. Opent the .sln file and build and go

Credits

This example was created by James George for Specular's project Curio Aquarium, which was financially supported and encouraged by the Kinect For Windows team.

ofxUI is by Reza Ali

ofxCv is by Kyle McDonald

ofxImageSegmentation by James George based on this example