Defect Detection Method for Substation Equipment Based on Lightweight Improved YOLOv5s Algorithm

Intelligent substation equipment defect detection task exists in the defect distribution, defects in the appearance of a variety of changes in the characteristics of the features, and feature extraction difficulties. This paper proposes a defect detection method for substation equipment based on the lightweight, improved YOLOv5s algorithm. The method takes YOLOv5s as the base detection model. It reconstructs the convolution operation process of the model to obtain data features based on the Ghost lightweight module, which improves the detection speed of the model. A loss function based on the KL scatter distribution is used as the target frame localization loss function to improve the model's accuracy for fuzzy image detection. The results show that the method proposed in this paper can improve the detection accuracy of the fuzzy fault target image while ensuring real-time detection, which meets the needs of intelligent substation equipment defect detection.