Multiple objects tracking based on multiple information integration

Multi-target tracking algorithm often fail when targets are covered, or move fast, and it cannot be recovered from the failure. To solve this problem, firstly we use multiple information which integrate the target motion information and shape information. Based on the Fisher Criteria, we make the distance between same targets as close as possible which the distance between different targets far away. Secondly, the single target tracker based on strong discriminative ability and the Kalman predictor can track accurately when the target is covered or moves fast. The experimental results show that our multi-target tracking algorithm can track target in occlusion or in fast moving accurately in real time.