A* algorithm with dynamic weights for multiple object tracking in video sequence

作者：Xi, Zhenghao^*; Xu, Dongmei; Song, Wanqing; Zheng, Yang

来源：Optik, 2015, 126(20): 2500-2507.

DOI：10.1016/j.ijleo.2015.06.020

• Multi-object tracking
• A* algorithm
• Flow network model
• Dynamic constraint

In the clutter environments, persistently tracking multiple objects is still a very challenging problem. In this paper, an A* algorithm with dynamic weights is developed to solve this problem. Firstly, the problem of multiple object tracking is considered as an integer programming of flow network model. Based on this model, afterward, we relax the problem of integer programming problem to a standard linear programming and achieve the global optimal solution with the proposed algorithm, which has been proven in this work. Compared with other advanced methods employed in the modern complex environments, the proposed method is on the merit of lower calculation complexity and better tracking accuracy and robustness. Finally, the simulation result is demonstrated, which reveals that the proposed algorithm saves the time costs vastly.

• 出版日期2015
• 单位北京科技大学; 上海工程技术大学