基于双光识别和YOLOv5模型的火灾风险分析与智能预警系统设计  

Fire risk analysis and intelligent early warning system design based on dual light identification and YOLOv5 model

作　　者：胡华斌 李叙达 杨运涵 HU Huabin;LI Xuda;YANG Yunhan

机构地区：[1]江西核电有限公司,江西九江332000

出　　处：《电力系统装备》2025年第1期111-114,共4页Electric Power System Equipment

摘　　要：发电站作为重要的能源基础设施,其火灾风险具有隐蔽性和破坏性,一旦发生火灾将造成严重的经济损失和安全威胁.为了提高发电站火灾预警的准确性与及时性,文章设计了一种基于双光识别和YOLOv5模型的火灾风险分析与智能预警系统.该系统结合红外光与可见光两种检测手段,通过YOLOv5模型对监控画面中的火灾特征进行实时识别与检测.研究结果表明,该系统在复杂发电站环境中具备较高的火灾识别准确率和响应速度,为发电站火灾防控提供了一种高效、智能的解决方案.Power station as an important energy infrastructure,its fire risk is hidden and destructive,once the fire will cause serious economic losses and security threats.In order to improve the accuracy and timeliness of power station fire warning,a fire risk analysis and intelligent warning system based on dual light recognition and YOLOv5 model is designed in this paper.The system combines infrared light and visible light detection methods,and uses YOLOv5 model to identify and detect fire characteristics in the monitoring screen in real time.The results show that the system has high fire identification accuracy and response speed in complex power station environment,and provides an efficient and intelligent solution for power station fire prevention and control.

关 键 词：双光识别技术 发电站火灾预警 YOLOv5 火灾风险分析 智能预警系统 

分 类 号：TM08[电气工程—电工理论与新技术]