Temporal Recursive Pedestrian Trajectory Identification for Multi-camera Surveillance System

Trajectory identification is fundamental and widely used in intelligent video surveillance systems. In this paper, a temporal recursive method is proposed to analyze and determine the trajectories of pedestrians that are monitored under a multi-camera network. The trajectory provides the information on pedestrian´s transition among cameras from a macroscopic perspective. Region´s topological structure and camera observations are exploited for the determination of trajectory. After that, the uniqueness of the solution is discussed and the conditions for multiple solutions are disclosed. Experiments on surveillance video database of Shanghai World EXPO 2010 and comparison with previous work are presented to demonstrate the effectiveness and advantages of the proposed recursive algorithm for trajectory identification.