首页|期刊导航|现代信息科技|基于机器视觉的口岸尾随行为检测方法研究

基于机器视觉的口岸尾随行为检测方法研究OA

Research on Port Tailgating Behavior Detection Method Based on Machine Vision

中文摘要英文摘要

为了检测不法分子在自助通关闸机前尾随合法人员的行为,文章采用了一种通过目标检测与目标跟踪方法分析行人运动属性特征的方案,对尾随行为的检测进行了研究.首先结合交叉学科内容对该行为进行定义,再采用目标检测算法识别出其中的通关人体.利用目标跟踪算法提取目标的运动轨迹,获取可疑人员与合法通关人员的空间位置关系.对目标的运动属性进行定量研究、设定阈值,分析人体的运动特征,得到尾随判定结果.由此对尾随行为进行表征,构建出特定场景下尾随行为的自动识别模型.实验结果表明,算法能够有效检测边检大厅尾随人员,实现区域内人员风险行为的安全管控.

To detect the behavior of unauthorized individuals tailgating legitimate passengers at self-service border control gates,this paper studies the detection of tailgating behavior using a scheme of analyzing pedestrian motion attribute characteristics by Target Detection and target tracking method.First,the behavior is defined through an interdisciplinary content.Then,a Target Detection algorithm is used to identify the legitimate individuals passing through the gates.A target tracking algorithm is applied to extract the movement trajectories of the subjects,obtaining the spatial relationship between suspicious individuals and legitimate passengers.The movement attributes of the targets are quantitatively studied,the threshold is set,the movement characteristics of individuals are analyzed,and the follow-up judgment results are obtained.This enables the representation of tailgating behavior and the construction of an automatic recognition model for tailgating behavior in specific scenarios.Experimental results show that this algorithm can effectively detect tailgaters in border inspection halls and achieve safe management and control of risky behaviors within the area.

李润涵;张梅

中国人民警察大学 研究生院,河北 廊坊 065000中国人民警察大学 移民管理学院,河北 廊坊 065000

计算机与自动化

机器视觉目标检测运动分析异常行为

Machine VisionTarget Detectionmovement analysisabnormal behavior

《现代信息科技》 2025 (12)

63-67,73,6

国家重点研发计划重点专项(2023YFC3321600)中国人民警察大学研究生科技创新计划项目(YJSKC2511)

10.19850/j.cnki.2096-4706.2025.12.013

评论