基于MeanShift与Group Search Optimizer的视频目标跟踪方法

摘要

针对传统的Meanshift方法在目标快速运动的情况下，目标跟踪易丢失问题，提出了一种将Meanshift与Group Search Optimizer融合的视频目标跟踪方法。该方法可以对运动目标位置进行全局搜索，根据Meanshift算法的跟踪结果判断是否要进行Group Search Optimizer算法优化，并且在目标跟踪丢失时可以快速重新找回原目标进行跟踪，提高了目标跟踪的鲁棒性。实验结果表明该方法可以有效提高对快速运动目标跟踪的准确性，具有较好的鲁棒性。%According to the traditional Meanshift method in target tracking lost problems under fast moving conditions,a video target tracking method of Meanshift combined Group Search Optimizer is proposed in this paper.This method can search the position of moving target in the global scope.According to the tracking result of Meanshift algorithm to decide whether carry out Group Search Optimizer algorithm.The method can fast find back the original targets to track again once losing then during a long period tracking,and improve the robustness of tracking.