Defect Detection Method for Substation Insulators and Overhead Equipment Based on an Improved YOLOv10

The safe and stable operation of substations is crucial to power systems. Traditional manual defect detection methods have limitations such as strong subjectivity, high missed detection rates, and low efficiency. Existing YOLO series algorithms have insufficient feature capture capability in complex substation scenarios and have difficulty distinguishing similar defects. Therefore, this paper proposes a defect detection method for substation equipment based on an improved YOLOv10 algorithm. Based on YOLOv10, the method adopts multi-scale convolution and multi-cooperative attention mechanisms to reconstruct the C2f module in the neck module of the detection layer. The improved C2f module is used to optimize the feature extraction algorithm and enhance the ability of the algorithm to identify targets at different scales. Simulation experiments verify the effectiveness and practicality of the proposed method.