舰船电子工程2019，Vol.39Issue(8)：51-55,5.DOI:10.3969/j.issn.1672-9730.2019.08.013
基于核相关性滤波的长时间目标跟踪方法
Long-time Target Tracking Method Based on Kernel Correlation Filtering
摘要
Abstract
As the tracking target is easy to fail when it changes rapidly and the target can not be retracked after lost,a long-time target tracking algorithm based on kernel correlation filtering is proposed. The algorithm introduces the idea of multi-scale detection and searches the optimal scale of the tracking target,which effectively improves the robustness of the tracker when the target changes rapidly. At the same time,TLD algorithm framework is introduced,detection module and learning module are added,target is re-locked when the tracker fails to track,and extended Kalman filter is introduced to predict the detection scan?ning area,which can improve the speed and accuracy of the algorithm. The experimental results show that the long-time target track?ing algorithm based on correlation filtering can achieve stable long-time target tracking.关键词
核相关性滤波/TLD/目标跟踪/多尺度检测Key words
kernel correlation filtering/TLD/target tracking/multi-scale detection分类
信息技术与安全科学引用本文复制引用
陈均瑞,盛守照,李郴荣..基于核相关性滤波的长时间目标跟踪方法[J].舰船电子工程,2019,39(8):51-55,5.基金项目
航空基金项目(编号:2016ZC52018)资助. （编号:2016ZC52018）
