摘要
近些年,随着轨道交通运输网络迅猛发展,地铁向着更加安全、便捷、舒适方向发展。准确掌握地铁车厢内实际人数,并在地铁站台显示出相关信息,不仅可以提高乘客乘车的舒适感和便捷性,同时缓解地铁站台固定区域附近乘客拥挤的情况。本文采用VGG16作为特征提取模块,采用特征金字塔构建密度图。本系统利用现有地铁视频监控信息,对地铁车厢人员密度进行监测,再将密度信息传输至乘客信息服务系统,并保持屏幕上多媒体信息、安全提示信息及其发布信息不变,帮助乘客及地铁运营管理者实时了解地铁车厢内拥挤情况。
In recent years,with the rapid development of the railway transport network,safety,convenience and comfort are more concerned in the subway system maintenance.Accurate crowd counting and information distribution on the station platform can not only reduce the crowd congestion but also improve the customer satisfaction.This project employs an computer vision based scheme to do the crowd prediction job.Specifically,the submodule of the VGG-16 model is chosen as the feature extractor and a simplified version of the feature pyramid network is used as the decoder to recover the crowd density map.Images of the deployed surveillance cameras are feed into the deep learning algorithm and the results are transmitted to the passenger service information system in real time.All the multimedia information including the camera scenes with the crowd counting,safety notice and other information is displayed on the screen to help the passengers to understand the situation.
作者
陈享成
Chen Xiangcheng(School of Electronic Engineering,Zhengzhou Railway Vocational and Technical College,Zhengzhou Henan,450052)
出处
《电子测试》
2022年第13期64-66,26,共4页
Electronic Test
基金
河南省重点研发与推广专项“基于机器视觉技术的地铁车厢拥挤度提示系统研究与应用(202102310525)”。
关键词
机器视觉
拥挤度
地铁
Computer Vision
Crowd Counting
Subway system
