土木工程与管理学报2024，Vol.41Issue(3)：1-11,11.DOI:10.13579/j.cnki.2095-0985.2024.20240064
基于计算机视觉的临边洞口坠落事故自动预警方法
An Automatic Warning Method of Falling from Edges and Openings Based on Computer Vision
摘要
Abstract
As a major hazardous source of falling accident from height,edges and openings have al-ways been the focus of safety management on construction sites.However,the complexity of construc-tion sites and a large number of edges and openings make managers difficult to monitor them simulta-neously.Therefore,this study proposed an automatic warning method for falling from edges and open-ings based on computer vision,to assist managers in mastering on-site information and prevent safety accidents in advance.Firstly,videos are captured by cameras installed on tower cranes and used to i-dentify and locate site the workers.Then,the location of hazardous area is determined,and the dis-tance between the workers and the hazardous area is measured automatically.Finally,the workers'safety state is judged and warning information is generated in a hierarchical way according to the es-tablished judgment rules.The proposed method was applied in Terminal 2 of Xinqiao International Airport.The results show that the overall measurement error is within 6％,which can effectively judge worker's safety state and assist in safety management.关键词
安全管理/高处坠落/临边洞口/安全预警Key words
safety management/falling from height/edge and opening/safety warning分类
信息技术与安全科学引用本文复制引用
李建华,陈宋春,张可嘉,郭红领,骆杰,方东平..基于计算机视觉的临边洞口坠落事故自动预警方法[J].土木工程与管理学报,2024,41(3):1-11,11.基金项目
国家自然科学基金(52278310) （52278310）
