YOLOv4算法在加油站卸油人员离岗检测中的应用研究  

Research on application of YOLOv4 algorithm in off-duty detection of unloading personnel in gas sta tions

作　　者：武熠明 徐明智 幸贞雄 周远生 姚宗泉 WU Yiming;XU Mingzhi;XING Zhenxiong;ZHOU Yuansheng;YAO Zongquan(Guizhou Institute of Labor Protection Science and Technology,Zunyi Guizhou 563000,China)

机构地区：[1]贵州省劳动保护科学技术研究院,贵州遵义563000 [2]贵阳红鸟智能技术服务有限公司,贵州贵阳550014

出　　处：《工业安全与环保》2023年第6期11-14,共4页Industrial Safety and Environmental Protection

基　　金：贵州省科研机构创新能力建设专项资金(黔科合服企[2020]4014)。

摘　　要：为督促卸油作业现场监护人员履职尽责,坚守岗位,防止生产安全事故发生,提出一种基于YOLOv4机器学习算法的卸油人员离岗检测方法,通过实时检测卸油口状态和现场作业人员数量,判断人员是否离岗。对检测方法进行测试,结果显示,该方法能够准确检测卸油口状态、追踪卸油过程中的现场作业人员,能够较好地检测到人员离岗行为。In order to supervise the guardians at oil unloading operation site to perform their duties and adhere to their posts and prevent production safety accidents,a YOLOv4 machine-learning algorithm-based oil unloading personnel departure detection method is proposed to determine whether personnel leave their posts by detecting the status of oil unloading port and the number of field workers in real time.The results show that the method can accurately detect the status of the unloading port,track the field workers in the process of oil unloading,and better detect the personnel leav ing the job.

关 键 词：安全管理 加油站 智能监测 人员离岗 

分 类 号：U473.8[机械工程—车辆工程] TP391.41[交通运输工程—载运工具运用工程] TE88[交通运输工程—道路与铁道工程]