Insulator defect detection algorithm based on improved YOLOv8

Insulators, as critical components in power transmission lines, play a vital role in ensuring the safe operation of power systems. Existing object detection algorithms often suffer from high false negative rates and insufficient accuracy in identifying small objects in complex environments. To address these issues, this paper proposes an improved YOLOv8-based insulator defect detection algorithm. First, the DSC-C2f module is used in the backbone network to replace the original C2f module, enhancing the extraction of scratch features and constructing a feature extraction module tailored to the characteristics of insulator cracks, thereby improving the ability to extract features of insulator cracks; then, the CPCA attention mechanism is introduced to enhance the backbone network's ability to extract features of objects with varying shapes, thereby improving detection accuracy; Finally, the loss function is replaced with the DIOU loss function, which better guides the model toward the center point of the bounding box, especially when the IoU of the bounding box is low, thereby improving localization accuracy. Experimental data show that the improved algorithm achieves a 1.4% and 6.8% improvement in mAP50% and mAP50%:90%, respectively, compared to the original algorithm, meeting the real-time and accuracy requirements for insulator inspection.