A method for Multi-Cameras Human Tracking Based on Reverse Epipolar Line Geometry

In human body motion tracking research, both the tracking precision and self-occlusion are difficult parts for researchers. This paper proposes a multi-camera human-tracking method on the basis of reverse epipolar Geometry, and this method unifies reverse epipolar Geometry,the data association,the multi-view geometry relations,the Kalman filtering and epipolar Geometry equation and etc. to track the motion path of human body. The experiment indicates that this method can track body motion effectively and obtain 3D reconstruction and has solved the self-occlusion to some extent.