基于算法融合的红外成像目标稳健跟踪方法

针对复杂背景下的红外图像机动目标跟踪问题,在对现有的多种红外成像目标跟踪算法进行分析与比较的基础上,提出了一种基于算法融合的红外成像目标稳健跟踪方法。该方法在对各跟踪器输出的目标位置测量值序列采用基于“当前”统计模型的模糊交互多模方法进行处理的基础上,采用基于总均方误差最小规则的自适应加权融合方法对目标状态的多个滤波与预测值进行综合处理,较大程度上提高了系统的跟踪精度与稳定性。仿真结果显示了该方法的有效性与稳健性。

Method for IR imaging target robust tracking based on algorithm fusion

Aiming at the problem of IR imaging moving target tracking under a complex background,a method for robust IR imaging target tracking based on algorithm fusion is proposed on the basis of analyzing the existing algorithms for IR imaging target tracking.By processing the observed data sequences of target position using the fuzzy IMM approach based on current statistics model,the proposed method integrates multiple filtering and predicted values of target state from different trackers using the adaptive weighting fusion approach based on total least-squares error rule,which improves the precision and robustness of tracking system. The experimental results show the effectiveness and robustness of the proposed method.