2023，59（6）改进YOLOv7与DeepSORT的佩戴口罩行人跟踪赵元龙1，单玉刚2，袁杰11.新疆大学电气工程学院，乌鲁木齐8300172.湖北文理学院教育学院，湖北襄阳441053摘要：针对视频序列中因脸部遮挡、漏检而造成的无法正确判断行人是否佩戴口罩的问题，提出一种基于改进YOLOv7与DeepSORT的佩戴口罩行人跟踪算法。该算法将口罩检测、行人检测与跟踪相结合，通过在YOLOv7的主干网络中添加注意力机制，增加浅层特征图，加强网络对小目标的感知能力，提高口罩检测与行人检测精度；帧内关系模块利用匈牙利算法进行帧内目标关联，对行人进行口罩佩戴标记；将方向差因素加入到DeepSORT算法的关联代价中，消除跟踪轨迹的历史预测方向和新检测速度方向不一致问题；使用改进的DeepSORT算法对行人进行跟踪，并对每条轨迹进行口罩佩戴标记更新，实现对佩戴口罩与未佩戴口罩行人的跟踪。实验结果表明，改进的YOLOv7网络平均检测精度mAP50相比原始算法提升了3.83个百分点；在MOT16数据集上，该算法的跟踪准确性MOTA相较DeepSORT算法提高了17.1个百分点，跟踪精度MOTP提高了2.6个百分点。与检测算法相比，提出的算法能够跟踪到更多的行人是否佩戴了口罩，具有更好的效果。关键词：行人跟踪；口罩检测；注意力机制文献标志码：A中图分类号：TP391.4doi：10.3778/j.issn.1002-8331.2210-0479WearingMaskPedestrianTrackingBasedonImprovedYOLOv7andDeepSORTZHAOYuanlong1,SHANYugang2,YUANJie11.SchoolofElectricalEngineering,XinjiangUniversity,Urumqi830017,China2.SchoolofEducation,HubeiUniversityofArtsandScience,Xiangyang,Hubei441053,ChinaAbstract：ApedestriantrackingalgorithmbasedonimprovedYOLOv7andDeepSORTisproposedtosolvetheproblemthatwhetherpedestrianswearmaskscannotbecorrectlyjudgedduetofaceocclusionandmisseddetectioninvideosequences.Thealgorithmcombinesmaskdetection,pedestriandetectionandtracking.Firstly,byaddingattentionmecha-nismtothebackbonenetworkofYOLOv7,shallowfeaturemapsareaddedtoenhancethenetwork’sabilitytoperceivesmalltargetsandimprovetheaccuracyofmaskdetectionandpedestriandetection.Secondly,theintra-framerelationshipmoduleusestheHungarianalgorithmtocorrelatetheintra-frametargetsandmarkthemaskwearingofpedestrians.Then,thedirectiondifferencefactorisaddedtotheassociationcostoftheD...
