红外点目标跟踪方法综述

Review of Point Target Tracking Methods Based on Infrared Sensor

综述了常见的红外点目标跟踪方法。介绍了全局最近邻(GNN)、概率数据关联(PDA)、联合概率数据关联(JPDA)、多假设跟踪(MHT)和动态规划(DPA)等算法,讨论了它们的区别、联系。介绍了概率假设密度(PHD)滤波方法,指出它在跟踪多目标方面相比其他算法的优势。展望了红外点目标跟踪方法的研究前景,提出对DPA和PHD两算法的继续研究是今后的重要研究方向。

Common methods are reviewed to detect and track point targets. During the review of con- ventional tracking methods, such as global nearest neighbor （GNN）, probabilistic data association （PDA） , joint probabilistic data association （JPDA） , multiple-hypothesis tracking （MHT） , and dynam- ic programming algorithm （DPA） , the differences and connections are discussed. The probabilistic hy- pothesis density （PHD） filter for multi-target tracking is also mentioned, and its superiority to other methods in performance is put forward. It is indicated that the improvements of DPA and PHD are possi- ble research areas for further study.