计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (5): 242-246.DOI: 10.3778/j.issn.1002-8331.1609-0233

• 工程与应用 • 上一篇    下一篇

基于双目视觉的接触网磨耗在线检测研究

王延华,李  腾   

  1. 北京交通大学 电气工程学院,北京 100044
  • 出版日期:2018-03-01 发布日期:2018-03-13

Research on on-line measurement of contact wire wear based on binocular vision

WANG Yanhua, LI Teng   

  1. School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044, China
  • Online:2018-03-01 Published:2018-03-13

摘要: 针对电力接触线磨耗检测中效率低、实时性差和精度低的问题,设计了一种基于线阵相机的非接触式测量系统。基于双目立体视觉原理和各坐标系关系,推导了接触线的磨耗表达式。系统采用递归逼近高斯滤波方法对采集到的图像进行滤波再进行边缘点提取,然后转化为磨耗长度并实时显示,并利用坐标变换对振动状态下测量的磨耗进行了误差分析。实验表明,能够实时地检测出电力接触线的磨耗大小,测量误差小于0.55 mm。

关键词: 接触线磨耗, 线阵, 双目立体视觉, 递归逼近高斯滤波, 振动

Abstract: To overcome the shortcomings of low efficiency and accuracy of contact wire wear detection, a non-contact measurement system based on linear array camera is designed. Based on the principle of binocular stereo vision and the conversion between the coordinate systems, it can derive the formula of contact wire wear. The system uses infinite impulse response filter to filter the images and then extracts the edge pixels points which can be converted to the distance and displayed on the MFC. It also analyzes the error caused in the vibrating state by using the coordinate transformation. Experiments show that the wear of the power contact wire can be measured accurately and the measurement error is less than 0.55 mm.

Key words: contact wire wear, linear array camera, binocular vision, recursive implementation of Gaussian filter, vibrate