Intelligent detection method of pontoon off-line of cold source interception network based on YOLOv5

Aiming at the problem of the drop in the carrying capacity of the interception net caused by the disconnection of the cold source interception net buoy at the water intake of the nuclear power plant, which affects the safety of cold source water intake. Based on the YOLOv5 target detection technology, this research proposes an intelligent detection method for cold source interception net buoy disconnection, which helps to grasp the position and status of the interception net buoy in real time, and put forward corresponding treatment measures in time. This method is based on deep learning and traditional vision technology, obtains the position of the buoy through the YOLOv5 target detection model, and classifies and judges its status, and detects the abnormality of the early warning buoy out of the network. This method is independently developed based on deep learning networks such as YOLOv5. The test results show that the recall rate of the floater off-net recognition is over 95%. This method can effectively detect the abnormal state of the buoy of the cold source interception network at the water intake of the nuclear power plant, and provides a certain safety guarantee for the cold source water intake of the nuclear power plant.