一种贝叶斯检测跟踪阈值确定方法

A New Method of Choosing Detection Threshold in Bayesian Coupling of Detection and Tracking

利用贝叶斯后验概率方法对目标进行检测和跟踪处理时,在分析有目标和没有目标情景下后验概率分布的基础上,文中给出了一种确定检测阈值的方法。利用该方法确定的检测阈值,能够利用有目标和没有目标情景下的后验概率分布,在对系统检测和跟踪性能影响很小的情况下,有效地控制系统的虚警率。

When coupling of detection and tracking of target is processed by using Bayes theory, based on the difference of posterior probability distribution of existing target between the situation of target being present and the situation of target being absent, a new method of choosing threshold of detection is proposed in this paper. The new method of choosing threshold of detection makes use of the difference of posterior probability distribution of existing target between the situation of target being present and the situation of target being absent. Thus, it can lower the false alarm rate of system, and has small influence on the performance of detection and tracking of system.