基于局部相似的红外小目标跟踪算法

连续图片序列中的运动对象与其局部背景保持相对稳定。相邻图像序列的目标局部灰度分布具有相似性。针对红外小目标跟踪杂波干扰与运动模糊问题,设计了基于局部相似的目标增强方法,并提出一种基于局部相似和运动估计的跟踪算法。该算法根据目标的局部相似性构建搜索空间,采用运动估计机制缩小搜索域,然后利用时空上下文学习跟踪算法产生跟踪结果。实验表明,本文方法能够有效应对杂波干扰和运动模糊情况,实现红外小目标的稳健跟踪,且具有较好的实时性。

Infrared dim-small target tracking algorithm based on local similarity

The moving objects in the continuous picture sequence and their local background remain relatively stable.The local pixel distribution of the target between adjacent frames is similar.Aiming at the problem of clutter interference and motion blur in infrared small target tracking,a target enhancement method based on local similarity is designed,a tracking algorithm based on local similarity and motion estimation is proposed.The improved algorithm constructs the search space according to the local similarity of the target,uses the motion estimation mechanism to narrow the search area,and then uses the spatio-temporal context learning tracking algorithm to generate the tracking results.Experiments show that the method in this paper can effectively deal with clutter interference and motion blur,achieve robust tracking of small infrared targets,and has good real-time performance.