高压线路巡线机器人视觉导航中的单目测距算法  

A monocular location algorithm for inspection robot of high-voltage transmission line

作　　者：程立[1,2] 吴功平[1] 

机构地区：[1]武汉大学动力不机械学院,湖北武汉430072 [2]中南民族大学计算机科学学院,湖北武汉430074

出　　处：《光电子．激光》2016年第9期941-948,共8页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金(61503418);国家民委科研(14ZNZ019)和;南民族大学中央高校基金(CZY16005)资助项目

摘　　要：为准确快速地定位高压线路上的障碍物,实现有效越障,提出了一种用于巡线机器人视觉导航的单目测距算法。采用Hough变换在视频帧图像中检测出机器人行走导线的两侧边缘;预先测取导线上可成像的最近点到镜头的距离且机器人移动过程中保持不变,由此处导线两侧边缘直线纵坐标差和障碍物处的两直线纵坐标差之比,得到障碍物到镜头的距离并对其进行线性修正,最终可得到障碍物沿导线到镜头的距离。试验结果表明,提出的方法在5 000mm内测量误差小于5%,可测距离至少能达7.4m;且算法具有精度较高、速度快、鲁棒性强、需测量参数较少和易于实现等优点。In order to locate obstacles on high-voltage transmission line fast and accura tely and cross them effectively,an algorithm for monocular vision distance measurement in inspection robot is proposed.First,according to position and pose of camera to the transmission line,a rough estimation of dip angle of the transm ission line in images from video frame is acquired.And then the parameters of both edges of the transmission line are detected fast from coarse to fine in which down-sampling and two-stage Hough transformation are used based on the estimated angle.At last,th e distance between the nearest point on the line in image and camera lens can be measured in advance,and according to the ratio of d ifference of vertical axis on the edges line of the point to that of the positio n of the obstacle,the distance between the obstacle and camera lens can be gaine dthen ,after revising,the distance along the transmission line between these t wo positions used in navigation for ro bot can be gained.Tests results show that the algorithm can achieve such accuracy as measurement error less than 5% within 5000mm.At the same time,the running speed of the algorithm is very fast and the processing time is less than 20ms.

关 键 词：巡线机器人 单目定位 视觉导航 HOUGH变换 针孔模型 越障 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]