基于Transformer与信息融合的绝缘子缺陷检测方法

Insulator defect detection method based on Transformer and information fusion

针对现有绝缘子航拍图像背景复杂,闪络、破损缺陷检测困难的问题,本文提出一种全局与局部信息融合(GLIF)-YOLOv8s绝缘子缺陷检测算法。该算法采用EfficientFormerV2作为主干网络,以提高模型对全局信息的提取能力;基于全局与局部信息设计特征增强模块,通过信息融合减少深层网络信息的丢失。在绝缘子缺陷数据集上进行消融实验与对比实验,结果表明:本文算法在绝缘子缺陷数据集上的平均精度均值为91.6%,其对闪络和破损缺陷的检测平均精度分别达到82.3%和92.9%;与其他主流算法相比,本文算法的检测框置信度更高。

Aiming at the existing insulator aerial images,which have complex backgrounds and are difficult to detect flashover and broken defects,a global and local information fusion(GLIF)-you only look once v8s(YOLOv8s)insulator detection algorithm is proposed.The algorithm uses EfficientFormerV2 as the backbone network to improve the model’s ability to extract global information.A feature enhancement module(FEM)is designed based on global and local information to reduce the loss of deep network information through information fusion.Ablation experiments and comparison experiments are carried out on insulators defects dataset,and the experimental results show that the proposed algorithm achieves 77.5%class-wide average accuracy,and its flashover and broken defect detection accuracy reaches 67.7%and 73.5%.Compared with other mainstream algorithms,the detection frame confidence of the proposed algorithm is higher.