基于图像解析的无人天车安全预警系统  

A safety early warning system for unmanned overhead crane based on image analysis

作　　者：朱亦辰 王婧雯 宋媛媛 亢克松 丁泽浩 王燕伟 Zhu Yichen;Wang Jingwen;Song Yuanyuan;Kang Kesong;Ding Zehao;Wang Yanwei(School of Computer Science and Technology,Shandong University,Qingdao 266000,Shandong;Shandong Business Institute,Yantai 264670,Shandong;HBIS Digital Technology Co.,Ltd.,Shijiazhuang 050000,Hebei;Tangshan Huitang IOT Technology Co.,Ltd.,Tangshan 063000,Hebei)

机构地区：[1]山东大学计算机学院,山东青岛266000 [2]山东商务职业学院,山东烟台264670 [3]河钢数字技术股份有限公司,河北石家庄050000 [4]唐山惠唐物联科技有限公司,河北唐山063000

出　　处：《河北冶金》2025年第1期72-75,共4页Hebei Metallurgy

摘　　要：随着物联网、人工智能等信息技术的发展,钢铁企业正在向以智能工厂为载体的智能制造模式转型。其中,起重机与库区的智能化建设,成为智能工厂建设极具代表性的一项技术。通过应用无人天车可以减少人工操作的风险,提高生产效率。目前钢厂对于使用无人天车进行钢卷搬运已经相对成熟,然而在钢卷搬运过程中仍存在潜在安全问题。结合实时监控摄像头和机器视觉技术设计了一种无人天车安全预警系统。该系统不仅可以判断无人天车下方的工作区域上是否有工作人员逗留,防止人车碰撞事故,而且可以判断无人天车的夹钳在移运过程中是否夹紧钢卷,防止钢卷掉落造成生产安全事故。两种功能都是基于目标检测模型YOLOv5及相应的后处理。应用结果表明,该系统能够有效判断工作区域是否有人员逗留以及钢卷是否被夹紧,保障生产安全。With the development of information technologies such as the Internet of Things and artificial intelligence,steel enterprises are transitioning to an intelligent manufacturing mode centered around smart factories.The intelligent construction of cranes and storage areas has become a representative technology in smart factory development.By applying unmanned crane,the risks associated with manual operations can be reduced while improving production efficiency.Currently,steel plants have achieved relative maturity in using unmanned crane for coil handling.However,potential safety issues still exist during the coil handling process.This paper proposes a safety warning system for unmanned crane by combining real-time surveillance cameras and computer vision technology.The system can not only judge whether there is staff staying in the working area under the unmanned crane to prevent collision accidents,but also judge whether the clamp of the unmanned crane clamps the steel coil during transportation to prevent the coil from falling and causing production safety accidents.The two functions are based on the target detection model YOLOv5 and the corresponding post-processing.The application results show that the system can effectively judge whether people are staying in the working area and whether the steel coil is clamped,to ensure the safety of production.

关 键 词：图像解析 无人天车 机器视觉技术 YOLOv5 安全预警 

分 类 号：TF55[冶金工程—钢铁冶金]