地铁列车清客遗留检测系统应用研究  

Application research on the remaining items detection system for cleared metro train

作　　者：刘琦 Liu Qi

机构地区：[1]北京市地铁运营有限公司,北京100044

出　　处：《现代城市轨道交通》2023年第6期99-104,共6页Modern Urban Transit

基　　金：北京市地铁运营有限公司科研项目——首都机场线车厢综合感知监控系统应用示范研究(2021000501081002)。

摘　　要：检查车厢是否清空是地铁运营的重要环节,自动化该流程可提高检查车厢的速度,保证地铁的运营效率。遗留物检测作为自动化清客系统的一个模块,主要负责检测场景中的遗留行李、箱包或其他未知类别物品,并将遗留物的位置标记在画面中。文章提出一种新的背景更新策略获得空车背景图像,用变化检测模型对比空车和清客时的背景,再由显著性检测模型进行二次确认,最终输出遗留物的区域和类别。该清客遗留检测系统可迅速定位车厢主要位置的遗留物品和人员,节约人力的同时提高地铁运营效率。Checking whether the carriages are empty is a crucial aspect of rail transit.Automating this process can expedite carriage inspection and enhance the operational efficiency of rail transit.As a module of the automated train clearing system,forgotten things detection is primarily responsible for identifying remaining luggage,bags,or other unidentifi ed items in the scene and marking their location on the screen.This module employs a new background update strategy to capture the background image of a passenger compartment.It utilizes a change detection model to compare the background of the empty passenger compartment with the cleared trains,followed by a signifi cance detection model for secondary confi rmation.Finally,it outputs the area and category of the forgotten things.The forgotten things detection system for cleared metro train can quickly locate the remaining items and passengers in the most positions inside the carriage,thereby saving manpower and improving metro operation effi ciency.

关 键 词：地铁 遗留物检测 背景建模 显著性检测 变化检测 合成数据 清客 

分 类 号：U231.92[交通运输工程—道路与铁道工程]