Research on Image Recognition Algorithm of Slag Accumulation in Slag Well Based on ShuffleNet and Improved YOLOv5

The normal operation of slag well can reduce the equipment failure and downtime caused by improper slag treatment, reduce environmental pollution and protect the ecological environment. However, the problem of slag accumulation in slag Wells occurs frequently, which seriously affects the operation efficiency and economic benefit of power plants. Aiming at the problem of slag accumulation in slag well, this paper improved YOLOV5s target detection algorithm by ShuffleNet, introduced the weighted bidirectional feature pyramid of context information and weight information for feature enhancement extraction, adopted position regression EIOU as a loss function to accelerate model convergence and improve detection performance, and verified the effectiveness of this algorithm through experiments. The results show that compared with the YOLOV5 algorithm, this algorithm can further improve the detection performance, has better detection effect, and the model is lighter. The application of this algorithm to the detection of slag accumulation in slag well can greatly improve the operation efficiency and economic benefits of the power plant, which is of great significance for the safety, environmental protection and economic operation of the power plant.