电厂作业人员安全装备穿戴检测系统

Safety equipment detection system for electric power construction worker

在电厂施工作业中,确保工人的安全排在首位,所以在作业区域内所有人员必须佩戴安全帽等安全装备。尽管各单位经常进行安全教育,甚至采取人力盯梢的方式进行监督,但总有心存侥幸者因为各种理由不能保证时刻佩戴安全帽。研究并设计了安全帽识别方法和系统,利用深度学习与计算机视觉技术,通过自动识别人员与安全帽等特征为安全员对现场监督提供有力保障。一方面,设计了基于YOLOV4的安全帽识别方法,能准确、实时检测施工人员是否穿戴了安全帽;另一方面,通过结合人脸识别和行人重识别技术有效识别未佩戴安全帽的工人的身份,形成实时告警,降低安全员的监管工作量。作业人员安全装备穿戴检测系统可为电厂施工的安全进行提供加强保障。

Safety helmets and other safety equipments must be worn by the workers who are in the power palnt construction area.Although various units often carry out safety education and even adopt the method of human stalking for supervision,some workers don’t like to wear helmets for various reasons.In this paper,a helmet recognition system which uses the advanced deep learning and computer vision technologies,was designed.the workers who don’t wear helmets can be recognized automatically.On one hand,a YOLV4-based helmet detection method was designed to detect the workers who don’t wear helmets accurately and real-time;On the other hand,face recognition and person re-identification technologies were combined to identify the workers who do not wear safety helmets.Operaror safety equipment detection system can strengthen the safety of power plant construction.