基于多特征与评估模型的红外小目标跟踪算法

Infrared small target tracking algorithm based on multi-feature and evaluation model

针对传统STC(时空上下文)算法的不足,在灰度特征的基础上,加入能表征局部区域灰度值变化程度的LWIE(局部加权灰度信息熵)特征,并利用红外目标与周围背景的局部对比度特征,对当前跟踪的每一个目标建立评估模型,以确定当前跟踪的目标准确有效.实验表明:该算法能准确地跟踪红外小目标,且具有良好的抗干扰性、实时性和鲁棒性.

In view of the shortcomings of traditional spatio-temporal context（STC）algorithm,on the basis of the gray features,the local weighted gray information entropy（LWIE）was joined to represent the variations of grayscale in local area.Then,the local contrast characteristics were used to establish an evaluation model for each tracking target to confirm whether the current tracking target was effective.Experimental results show that the proposed algorithm can track target effectively and accurately,which has good anti-interference performance,availability,real-time and robustness.