摘要
为解决目前人工处理分析无人机巡检图像效率低、检测结果受人为因素影响较大的问题,提出了一种用于检测绝缘子缺陷的改进YOLOv4故障检测模型。通过改进普通卷积算法以提升检测速度,使用数据增强方法提高YOLOv4对绝缘子缺陷检测性能,解决实际检测环境中缺陷图像数量少且识别精度低的问题。试验结果表明,所提方法的缺陷检测精度和召回率分别为0.91和0.96,能够满足电力线路绝缘子缺陷检测的鲁棒性和准确性要求。
In order to solve the problems of low efficiency of manual processing and analysis of UAV inspection images and great influence of human factors on detection results,an improved YOLOv4 fault detection model is proposed to detect insulator defects.Common convolution algorithm is improved to improve detection speed.The data enhancement method is used to improve the performance of YOLOv4 for insulator defect detection,and the problems of the scarcity of defect images and low recognition accuracy in the actual detection environment is solved.Experimental results show that the proposed method has a defect detection accuracy of 0.91 and a recall rate of 0.96,which can meet the requirements of robustness and accuracy of power line insulator defect detection.
作者
刘东东
LIU Dongdong(Fujian College of Water Conservancy and Electric Power,Yong′an 366000,China)
出处
《电工技术》
2022年第2期151-155,共5页
Electric Engineering
基金
2021年度福建水利电力职业技术学院教科研课题(编号YJRCKYQD202107)。