摘要
针对无人机(UAV)巡检输电线路采集玻璃绝缘子图像检测自爆缺陷问题,文中提出了一种基于图像分割技术的绝缘子自爆缺陷检测方法。通过RGB颜色通道将原始图像转化为灰度图像,采用三重分类Otsu方法将灰度图像中的像素按灰度值分为三类,结合Hough变换提取绝缘子串的主轴并将图像调整到水平方向。通过对绝缘子二值图像的连通区域进行标记,选定空间特征并建立自爆缺陷检测判别条件,利用连通区域的空间分布来锁定缺陷位置。实验结果表明:对100幅绝缘子自爆缺陷图像检测,检测精度为94.25%,平均检测时间仅为0.56s。
Based on the problem of self-explosion defect detection of glass insulator image collected by unmanned aerial vehicle(UAV)in power transmission line inspection,this paper proposes a self explosion defect detection method of insulator based on image segmentation technology.The original image is transformed into gray image through RGB color channel,and the pixels in gray image are divided into three categories according to gray value by triple classification Otsu method.Combined with Hough transform,the principal axis of insulator string is extracted and the image is adjusted to the horizontal direction.By marking the connected region of the binary image of insulator,the spatial features are selected and the self-explo sion defect detection and discrimination conditions are established.The defect location is locked by using the spatial distribution of the connected region.The experiment results show that the detection accuracy is 94.25%among 100 insulator self-explosion defect image,the average detection time of which is only 0.56s.
作者
沈杰
陈玉权
吴媚
张欣
孟悦
SHEN Jie;CHEN Yu-quan;WU Mei;ZHANG Xin;MENG Yue(Jiangsu Frontier Electric Technology Co.,Ltd.,Nanjing 211102,China)
出处
《信息技术》
2022年第5期165-170,176,共7页
Information Technology
关键词
图像分割
玻璃绝缘子
自爆
缺陷检测
故障定位
image segmentation
glass insulator
self breakdown
defect detection
fault location
