摘要
针对复杂环境中行人静动态监测的需求,设计了一个基于深度学习的行人检测与跟踪系统。文章首先利用YOLO系列算法实现对行人目标的检测,然后,在此检测结果的基础上有效结合SORT跟踪方法来实现对行人的多目标跟踪,并实现了易于用户操作的图形化系统界面。针对于无法平衡行人检测速度与精确度的问题,文章设计的系统提供了可以根据监测场景的不同而选择具有不同优点的算法模型的模型选择功能。通过实验表明,该系统能够较好地完成复杂环境下的行人检测与跟踪,推动行人检测与跟踪系统领域的进一步发展。
Aiming at the demand of pedestrian static and dynamic monitoring in complex environment,a pedestrian detection and tracking system based on deep learning is designed.In this paper,YOLO series algorithms are used to detect pedestrian targets.Then,based on the detection results,the SORT tracking method is effectively combined to achieve multi-target tracking of pedestrians,and a graphical system interface that is easy to operate is realized.Aiming at the problem that the speed and accuracy of pedestrian detection cannot be balanced,the system designed in this paper provides a model selection function that can select algorithm models with different advantages according to different monitoring scenarios.Experiments show that the system can better complete pedestrian detection and tracking in complex environments,and promote the further development of pedestrian detection and tracking system.
作者
张明娇
ZHANG Mingjiao(School of Rail Transportation,Soochow University,Suzhou,Jiangsu,China 215000)
出处
《长江信息通信》
2023年第9期19-21,共3页
Changjiang Information & Communications
