定向识别航拍绝缘子及其缺陷检测方法研究

针对现有绝缘子检测算法无法定向检测绝缘子及其缺陷的问题，提出了一种改进YOLOv5(you only look once v5,YOLOv5)算法的航拍绝缘子识别及其缺陷检测方法。通过定向标注航拍绝缘子图片，形成航拍绝缘子数据集和缺陷绝缘子数据集；在YOLOv5的主干特征提取网络引入轻量化注意力机制模块、在特征融合阶段使用改进的空间金字塔池化结构；通过改进YOLOv5网络的头部结构使其可以对绝缘子进行定向识别，并对损失函数添加角度损失分类。实验结果表明在检测时间由单张0.044 s到单张0.049 s并无显著增长的前提下，改进后的算法在测试集上的mAP(mean average precision)的值为95.00%,实现了定向识别绝缘子及其漏帽缺陷，还可应用到绝缘子视频流检测。为后续的绝缘子精确定位以及进一步故障检测打下良好基础。

Research on directional identification of aerial insulators andtheir defect detection methods

Aiming at the problem that the existing insulator detection algorithm cannot detect insulators and their defects in an orientedmanner, an aerial insulator identification and defect detection method improved by YOLOv5 algorithm is proposed. By orienting the aerialinsulator pictures, the aerial insulator dataset and defective insulator dataset are formed. The lightweight attention mechanism module isintroduced in the backbone feature extraction network of YOLOv5, and the improved spatial pyramid pooling structure is used in thefeature fusion stage. By improving the head structure of the YOLOv5 network, the network can perform directional identification ofinsulators and add angular loss classification to the loss function. The experimental results show that under the premise that the detectiontime does not increase significantly from 0. 044 s to 0. 049 s per sheet, the value of mAP (mean average precision) on the test set of theimproved algorithm is 95. 00%, which realizes directional identification of insulators and their leakage cap defects, and can also beapplied to insulator video stream detection. This provides a good basis for the subsequent precise positioning of insulators and furtherfault detection.