摘要
以提高复杂环境下地铁站点客流检测的有效性为目标,设计一种基于机器视觉的实时在线客流检测系统。确定客流检测的方案,明确摄像机具体的安装位置,分析其安装高度、摄像机视角与水平俯视角度之间的关系,以获取高质量的检测图像。对采集的客观图像进行背景提取与更新,运用背景帧差法对图像进行二值化处理,同时以形态学理论为基础对图像进行降噪处理,提取图像中更加完整的运动目标轮廓。通过对人物像素点的分析与统计,实现目标的跟踪,记载目标跟踪数据,统计客流数量,并在Python环境中开发了相应的客流检测软件系统。实验结果表明,本系统检测率高于90.00%。
In order to improve the effectiveness of passenger flow detection at subway stations in complex environment,a real-time online passenger flow detection system based on machine vision is designed.The scheme of passenger flow detection is determined,the specific location of camera installation is clarified,and the relationship between the installation height,camera angle and horizontal viewing angle is analyzed to ensure the quality of image acquisition.The background of the collected objective image is extracted and updated,and the image is binarized by background frame difference method.Meanwhile,the image is denoised on the basis of morphological theory to extract the more complete moving target contour in the image.Through the analysis and statistics of character pixels,the target tracking is realized with its data being recorded,the number of passenger flow is counted,and the corresponding passenger flow detection software system is developed in the Python environment.The experiment result shows that the detection rate of this system is higher than 90.00%.
作者
郑凯
王娟娟
谢国坤
王冠军
ZHENG Kai;WANG Juanjuan;XIE Guokun;WANG Guanjun(Xi’an Traffic Engineering Institute,Xi’an 710300,China)
出处
《机械制造与自动化》
2021年第4期156-158,186,共4页
Machine Building & Automation
基金
陕西省教育厅科研计划资助项目(20JK0743)。
关键词
机器视觉
地铁站点
客流检测
应用研究
machine vision
subway station
passenger flow detection
application research
