The model which predicts passenger distribution in a station using partial image sequence

In order that passengers can behave more spontaneously in a station, it is necessary to grasp the situation of the whole station and offer better service which suit various behaviors. Now, there is no method of grasping the situation everywhere in a station which is a complicated space. The human behavior in partial space can be grasped by using apparatus for exclusive use. We supposed that by using this partial grasps the situation of the whole station can be predicted in real time. The process is as follows. First, we acquired a passenger's position coordinate by the surveillance camera of the station yard. Next, we indicated movement of the passenger to the dead angle portion which has not been caught with a surveillance camera as a probability. Finally, we propose the model for grasping the state of the whole station yard. In thus appears that the situation of the whole station can be predicted in real time by the situation of partial space.