The fire prevention and control on the site of industrial production has always been an important part of enterprise safety management,as an important part of maintaining the stability of the national power system,the accurate prevention and control of fire accidents is of great significance.At present,deep learning technology has been widely used in various industries,and has shown great use value.Based on the YOLOv7 target detection model,this paper uses open data set data to build a smoke and flame target detection model applied to energy storage power stations.Adjusting the weight value of the loss function makes the model pay more attention to the detection of flame or smoke target.The experimental results show that the detection accuracy of smoke and flame targets is 85.88％and 76.03％,the detection recall rate of smoke and flame targets is 93.54％and 78.8％,and the mAP value of the model is 89.67％.The model has a good detection effect on the two types of targets.
关键词
火灾防控/深度学习/目标检测
Key words
fire prevention and control/deep learning/object detection
维普
万方数据