北京信息科技大学学报（自然科学版）2023，Vol.38Issue(6)：48-55,8.DOI:10.16508/j.cnki.11-5866/n.2023.06.007
多特征融合的抗遮挡相关滤波目标跟踪算法
Anti-occlusion correlation filtering target tracking algorithm based on multi-feature fusion
摘要
Abstract
To deal with the poor robustness of the traditional correlation filtering target tracking algorithm owing to background clutter,target scale change,target occlusion,a multi-feature fusion based anti-occlusion correlation filtering target tracking algorithm was proposed.The kernel correlation filter(KCF)algorithm was improved.By weighted fusing histogram of oriented gradient(HOG)features with multiscale block local binary pattern(MB-LBP)features,and secondary weighted fusing the obtained features with the features of the color name(CN),the feature expression ability was enhanced.A scaling filter was added for target scaling estimation.Peak sidelobe ratio(PSR)and target relocation algorithm were used to determine and detect the blocked target.OTB100 was selected as the data set,and the improved algorithm was compared with several mainstream algorithms in tracking effect.The experimental results show that the proposed algorithm has good robustness under the condition of background clutter,target scale change and target occlusion.关键词
相关滤波/目标跟踪/多特征融合/尺度滤波器/目标重定位Key words
correlation filtering/target tracking/multi-feature fusion/scale filter/target relocation分类
信息技术与安全科学引用本文复制引用
王慧炳,左云波,刘重阳..多特征融合的抗遮挡相关滤波目标跟踪算法[J].北京信息科技大学学报（自然科学版）,2023,38(6):48-55,8.基金项目
国家重点研发计划项目(2020YFB1713203) （2020YFB1713203）
