Abstract
Recently in-depth understanding of pedestrian behavior in public space is becoming significant with regard to achieving more sophisticated space design and flow control. Although there is a need to acquire automatically such behavior information, automatic human tracking is difficult under the situations that people move close to each other or are occluded by others. In this paper, we propose a new method of multiple human tracking under the complex situations. We review related works on both human tracking method by image processing and simulation model of pedestrian behavior, respectively. According to the review, we develop a method as a stochastic process, integrating them on the framework of general state space model. We apply the proposed method to the data acquired at the ticket gate of the railway station and confirm the high performance of the method. We also demonstrate the ability to acquire the OD data from the tracking result and show the future possibility.
