Transmission line insulator defect detection method based on improved YOLOv8

Aiming at the problem of insufficient detection accuracy of insulator defect images in transmission lines due to the complex background, the large difference in target size and the easy loss of small defect features, an insulator defect detection method based on improved YOLOv8 is proposed. Firstly, SENetV2(Squeeze-and-Excitation Network Version 2) multi-branch channel attention module is introduced into the backbone network of YOLOv8, which strengthens the defective feature expression and suppresses the complex background interference of insulator images through dynamic channel weight allocation; secondly, for the tiny defects Secondly, to address the problem of feature loss of tiny defects, a new high-resolution small target detection layer is added to the neck network, which integrates shallow localization details with deep semantic information to construct a cross-scale feature enhancement path to achieve accurate detection of tiny insulator defects. The experiments are based on the self-constructed insulator defect dataset, and the results show that the improved algorithm improves the precision by 3.4% and the recall by 3.9% compared with the original algorithm. The effectiveness of the method is verified by comparing it with other mainstream target detection algorithms.