基于改进YOLOv7模型的绝缘子缺陷检测研究

Research on Insulator Defect Detection Based on Improved YOLOv7 Model

绝缘子缺陷检测占据电力巡检的主导地位,传统的电力巡检方法需要人工实地勘察,缺乏安全性和实效性。为此,提出了一种改进YOLOv7模型的绝缘子缺陷检测算法,以提高绝缘子缺陷检测的精度和速度。该算法在YOLOv7的基础上引入了HorNet递归门控卷积重构目标检测颈部网络,解决了基网络缺乏全局建模、长距离建模能力的问题;采用SIoU Loss作为损失函数进而提高模型的收敛速度。实验结果表明,与基网络相比,该方法平均精度提高了2.8%,总平均精度均值达到了95.8%,相对于原始模型提高了4.1%,检测速度提高了53.3%,能够满足实时检测的要求。

Insulator defect detection occupies a dominant position in power inspection.However,traditional power inspection methods require manual field investigation and lack safety and effectiveness.Therefore,an improved insulator defect detection algorithm based on YOLOv7(you only look once)model is proposed to improve the accuracy and speed of insulator defect detection.Based on YOLOv7,the algorithm introduces HorNet Recursively Gated Convolution(GnConv)to reconstruct target detection neck network,which solves the problem that the base network lacks the ability of global modeling and long-distance modeling.SIoU Loss is used as a loss function to improve the convergence speed of the model.The experimental results show that compared with the base network,the Average Precision of the proposed method is increased by 2.8%,the total mean Average Precision(mAP)reaches 95.8%,which is increased by 4.1% compared with that of the original model and the detection speed is increased by 53.3%,which can meet the requirements of real-time detection.