April 1, 2012

During the last three weeks I was doing some research about "motion analysis in sport", "motion capturing" and espacially in new "3D-technolgies". I found a lot of helpful and interesting information. here a short list of the most important stuff I've found:

3D Motion Capturing

there are mainly three different approaches in 3d motion capture:

• suits: especially for motion capture designed suits with different sensors to get information about the joint movement of a athlete. (e.g http://www.3dsuit.com)
• marker based motion capture: the athlete has to be outfitted with several markers on different joints. multiple rgb cameras are tracking the position of this markers in order to reconstruct the movement in 3D
• marker free systems: there is an approach to match the images of several cameras to one 3D-Image without any markers.

There are now some new approaches in motion capturing with depth cameras. e.g:

• http://www.northkingdom.com/blog/motion-capture-in-3d-using-time-of-flight-camera/

http://www.google.ch/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=0CE0QFjAC&url=http%3A%2F%2Fai.stanford.edu%2F~koller%2FPapers%2FGanapathi%252Bal%3ACVPR10.pdf&ei=0BmAT-_FLMTrOf2_6JMH&usg=AFQjCNFUgvnSUiV59C_lpest5F-BClk1-g&sig2=j8m32dv84ylb8iigtsb4-A The interesting thing about this solution is, that one camera already is able to capture a 3D-Video even if it is not a full 360-view. Also, there are no markers needed since the depth information is enough to distinguish body parts from its surounding. One of the most popular example of this approach is the kinect camera from microsoft's Xbox. With a combination of a RGB- and a ToF-Camera, a real 3D-Video can be captured by the kinect. Since Kinect has a large OpenSource Community (openNI), there already are a lot of different libraries available.

3D Motion Analysis in Sport

I did not find a lot of Information about 3D Motion Analysis in sport. Therefore I assume that it is not much used yet. There are several reasons for it:

• framerate: the ToF-Cameras has not a very high framerate (around 30fps). For sport applications, this is not enough because it is not possible to have a good slow motion video.
• resolution: also, the resolution is not good enough.
• quite new: the new technology is not yet very famous and not a lot was implemented in the first years of its appearance.

Nevertheless, I see a great potential of this new technology. The depth camera gives you the needed information to analyse a lot of interesting things in a pretty easy way. For example the center of mass can be calculated and even the basic skeleton can be detected.