Automatic Classification of Substation Equipment Based on Multi-View Inspection Images

In addressing the automatic recognition and classification of equipment in the construction of smart substations, we propose a method based on multi-view inspection images for automatic classification of substation equipment. This method utilizes the YOLOv8 object detection algorithm for automatic classification of substation equipment. Through multi-view inspections of substation equipment to capture images from different angles, a segmented substation equipment sample set is used to train the YOLOv8 algorithm. The collected data is preprocessed, and the trained YOLOv8 algorithm is then used to classify and recognize images, achieving automatic classification of substation equipment. Experimental results demonstrate the successful detection of various types of substation equipment, including disconnectors, racks, insulating porcelain bottles, current transformers, surge arresters, and circuit breakers. At a threshold of 0.5, the average recognition accuracy for all classes is 0.502, indicating the application potential of the YOLOv8 algorithm in substation equipment recognition tasks. This method can provide certain technical support for the construction of smart substations.