摘要
为解决电子焊接实训过程由于电烙铁的错误放置、电源关闭不及时等行为容易引发火灾或烧伤事故的问题,对实训过程中的安全隐患进行了分析并制作了一套目标检测数据集,利用该数据集对骨干网不同的两种PP-YOLO目标检测模型进行了训练.经过对比,最终选择性价比更高的基于“MobileNetV3_large”骨干网的模型进行部署,形成了一套电子焊接实训装置安全报警系统.验证结果表明,该系统能够准确地检测安全隐患并发出报警提醒,可有效降低安全事故发生概率.
To solve the problem that fire or burn accidents are easily caused by the wrong placement of electric iron and untimely power supply shutdown in the process of electronic welding training,the potential safety hazards in the training process are analyzed,and a target detection data set is produced,which is used to train two PP-YOLO target detection models with different backbone networks.After comparison,the more cost-effective model based on the“MobileNetV3_large”backbone network is selected for deployment,and a safety alarm system for electronic welding training equipment is formed.The experimental results show that the system can accurately detect the hidden danger and send out an alarm,effectively reducing the probability of safety accidents.
作者
马军磊
栗伟周
张赞
MA Junlei;LI Weizhou;ZHANG Zan(Engineering Training Center,Xuchang University,Xuchang 461000,China)
出处
《许昌学院学报》
CAS
2023年第2期116-121,共6页
Journal of Xuchang University
基金
河南省高等教育教学改革研究与实践重点项目(2019SJGLX137)
许昌学院校级科研项目(2022YB038)。