中国电力 ›› 2019, Vol. 52 ›› Issue (4): 111-118.DOI: 10.11930/j.issn.1004-9649.201805103

• 信息与通信 • 上一篇    下一篇

高压输电线路巡检机器人障碍物识别与定位

张峰1,2, 郭锐1,2, 卢士彬1,2, 李振宇1,2, 杨波3, 孙晓斌3   

  1. 1. 国网山东省电力公司电力科学研究院, 山东 济南 250003;
    2. 山东鲁能智能技术有限公司, 山东 济南 250101;
    3. 国网山东省电力公司, 山东 济南 250001
  • 收稿日期:2018-05-10 修回日期:2018-07-28 出版日期:2019-04-05 发布日期:2019-04-16
  • 作者简介:张峰(1983-),男,高级工程师,从事电力机器人技术研究,E-mail:zhangf_0633@qq.com;郭锐(1978-),男,工学博士,高级工程师(教授级),从事电力特种作业机器人研究,E-mail:guoruihit@gmail.com;卢士彬(1987-),男,工程师,从事电力机器人控制与检测研究,E-mail:lushibin2009@qq.com
  • 基金资助:
    国家电网有限公司科技项目(520600160004)。

Obstacle Recognition and Location of Inspection Robot for High-Voltage Transmission Lines

ZHANG Feng1,2, GUO Rui1,2, LU Shibin1,2, LI Zhenyu1,2, YANG Bo3, SUN Xiaobin3   

  1. 1. State Grid Shandong Electric Power Research Institute, Jinan 250003, China;
    2. Shandong Luneng Intelligent Technology Co., Ltd., Jinan 250101, China;
    3. State Grid Shandong Electric Power Company, Jinan 250001, China
  • Received:2018-05-10 Revised:2018-07-28 Online:2019-04-05 Published:2019-04-16
  • Supported by:
    This work is supported by Science and Technology Project of SGCC (No.520600160004).

摘要: 根据高压输电线路地线上障碍物的几何形状和结构特点,提出了一种基于单目视觉识别的多传感器高压线路地线障碍物识别和测距定位的方法。该方法利用机器人本体上的摄像机获得图像,然后再根据障碍物的几何形状特征,判断出障碍物的类型。根据识别出来的障碍物特征点与摄像机的位置关系建立测距几何模型,从而计算出机器人到障碍物的距离。当机器人非常靠近障碍物或者正在通过障碍物时,摄像机测距定位方法不再适用,此时可以采用安装在机器人上的光电传感器和动态倾角仪,结合机器人的运行速度和传感器采集信号的变化,识别出障碍物的关键点,从而计算出机器人与障碍物的位置关系。

关键词: 机器人, 传感器, 高压输电线路, 障碍物识别与定位, 视觉识别

Abstract: According to the geometric shape and structural characteristics of the obstacles on the ground wire of high-voltage transmission lines, a method based on monocular visual recognition for the recognition and location of obstacles on multi-sensor high-voltage lines is proposed. Use the camera on the robot body to obtain the image, and then determine the type of the obstacle based on the geometry features of the obstacle. A distance measuring geometric model is established based on the recognized positional relationship between the obstacle feature point and the camera, thereby calculating the distance from the robot to the obstacle. When the robot is very close to obstacles or is passing through obstacles, the camera positioning method is not applicable. In this case, photoelectric sensors and dynamic inclinometers installed on the robot can be used in combination with the robot's running speed and the sensor's acquisition signal. The changes can identify the key points of the obstacle and calculate the specific position that has reached the obstacle.

Key words: robot, sensor, high-voltage transmission line, recognition and location of obstacles, visual identity

中图分类号: