论文部分内容阅读
本文论述了用红外图象序列跟踪目标的问题。证明库尔哈维开发的贝叶斯闭环估计算法很适于红外跟踪问题。红外焦平面阵列上的辐射强度模式适于用RSS的闭环表达式,在库尔哈维的算法中就是用这种方法对真正后验密度进行近似估计的。由RSS新建的公式可以导出对目标状态的估计。为了比较,使用以前开发的基于广义卡尔曼滤波器(EKF)的红外跟踪算法和基于RSS的新方法,借助红外图象序列跟踪目标。已证明,在EKF发散的高速情况中,RSS算法仍保持跟踪。
This article addresses the issue of tracking targets with infrared image sequences. It proves that the Bayesian closed-loop estimation algorithm developed by Kurhave is very suitable for infrared tracking. The radiation intensity pattern on the infrared focal plane array is suitable for the closed-loop expression of RSS, which is used in the algorithm of Kur’evii to estimate the approximate real posterior density. New formulas from RSS derive estimates of the state of the target. For comparison, the target was tracked with an infrared image sequence using a previously developed generalized Kalman filter (EKF) -based infrared tracking algorithm and a new RSS-based method. It has been shown that the RSS algorithm keeps track of the high-speed EKF divergence.