基于SNF的一种检测和跟踪性能优化方法

A Optimization Method of Detection and Tracking with SNF

对于一个监视系统,跟踪性能的优化要求检测系统和跟踪系统通过引入反馈进行联合设计,针对最大量测幅值邻域滤波器(SNF),提出了先验和后验检测跟踪优化准则,通过分析得出先验优化准则不可行这一结论,而利用后验优化准则同时对检测阈值和门值的联合优化,仿真结果表明,用后验优化准则进行检测和跟踪性能优化,不仅具有比恒虚警率更好的跟踪精度,而且其计算代价也大为降低。

In a surveillance system, it is required that detecting system and tracking system should be considered together to improve some filter performance. In this paper, the author investigates the prior and posterior optimization algorithms of the Strongest Neighbor Filter （SNF） respectively, and a conclusion is drawn that the prior optimization algorithm can not be implemented. Then both the gate size and the threshold are determined through the posterior algorithm. Simulation results show that the performance of the posterior optimization algorithm is superior to that of the constant false alarm probability, and the computation costs, furthermore, are reduced drastically.