摘要
为了解决目前人工巡检电力线路效率低、风险高的问题。提出了一种用于检测绝缘子缺陷的级联卷积神经网络结构。实验结果表明,文章方法的缺陷检测精度和召回率分别为0.93和0.97,能够满足高压输电线路绝缘子缺陷检测的精确性和检测效率的要求。
In order to solve the problems of low efficiency and high risk of manual inspection of power lines at present.This paper proposes a cascaded convolutional neural network architecture for detecting insulator defects.The experimental results show that the defect detection precision and recall rate of this method are 0.93 and 0.97,respectively,which can meet the requirements of the accuracy and detection efficiency of high-voltage transmission line insulator defect detection.
作者
刘东东
LIU Dong-dong(Fujian College of Water Conservancy and Electric Power,Yongan 366000,China)
出处
《电脑与信息技术》
2022年第3期15-18,共4页
Computer and Information Technology
基金
2021年度福建水利电力职业技术学院高层次人才项目(项目编号:YJRCKYQD202107)。
关键词
绝缘子故障
航拍图像
卷积神经网络
缺陷定位
故障检测
insulator failure
aerial imagery
convolutional neural network
defect location
fault detection
