Measurement of Pedestrian in Unknown 3D Environment Using Subtraction Stereo
Kenji Terabayashi, Yuki Hashimoto, Yuma Hoshikawa, Kazunori Umeda · 2009
This paper shows a method to measure pedestrians in unknown threedimensional environment for applications such as surveillance cameras. For this measurement, we have proposed a stereo camera system that focuses on moving regions. In the system, the correspondence problem of the stereo cameras is reduced by restricting the stereo matching process to moving regions. By focusing on moving regions, pedestrians can be detected easily based on standard labeling and information of size that is directly obtained from subtraction stereo without any given information about measurement environments. Tracking each pedestrian, the extended Kalman filter technique is employed to address occlusion problem usually seen in many people environment. The proposed method is verified by an experiment in unknown three-dimensional environment. 1. は じ め に