Application of lightweight YOLOv8n networks for insulator defect detection

To solve the problems of small insulator defect targets and complex background information in transmission lines, as well as the difficulty of edge-end devices to meet real-time detection requirements, a lightweight insulator defect detection algorithm based on YOLOv8n is proposed. The backbone network of YOLOv8n is reconstructed by introducing a lightweight bottleneck structure, GhostNetV2 BottleNeck, which reduces the number of network parameters and improves the detection speed of the model, and the CBAM attention mechanism is embedded in the backbone network, which improves the ability of the network to extract the target features, and thus improves the detection accuracy of the model. By validating the improved algorithmic model on the insulator dataset, the results show that the mean average accuracy of the improved algorithmic model reaches 85.7%, and the detection speed reaches 171.4 frames/s, which verifies the effectiveness of the improved algorithmic model for the detection of insulators and their defects.