基于机器视觉的尖轨轮廓检测方法  

Method for detecting sharp rail contour based on machine vision

作　　者：武福[1] 豆辉 武云 李忠学[1] 杨喜娟[3] WU Fu;DOU Hui;WU Yun;LI Zhongxue;YANG Xijuan(School of Mechatronic Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China;China Railway Guangzhou Bureau Group Co.Ltd.,Yongzhou Public Works Section,Yongzhou 425000,China;School of Eletronic and Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学机电工程学院,甘肃兰州730070 [2]中国铁路广州局集团有限公司永州工务段,湖南永州425000 [3]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《光学技术》2020年第4期453-460,共8页Optical Technique

基　　金：甘肃省重点研发计划项目(18YF1GD099);兰州市人才创新创业项目(2018-RC-107,2017-RC-82);甘肃省中小企业创新基金(18CX5JA014)。

摘　　要：针对尖轨相对于基本轨位置的轮廓检测问题,提出了利用线结构光辅助机器视觉的非接触式测量方法进行检测。在线结构光测量的基础上,建立了尖轨轮廓的三维检测模型,构造了测量尖轨与基本轨间的面差、间隙以及轨头宽度的数学模型,并采用Zhang-Suen与灰度重心法的融合算法对激光条纹进行提取和细化,最终得到了亚像素级的激光条纹中心线。本次实验在时速250km/h的18号道岔直线尖轨上进行测量,得到的数据与接触式测量的数据基本一致,验证了该方法的有效性和正确性,实现了对尖轨轮廓尺寸的非接触式检测。Aiming at the problem of contour detection of the position of the sharp rail relative to the basic rail,a non-contact measurement method using line structured light assisted machine vision is proposed for detection.On the basis of online structured light measurement,a three-dimensional detection model of the sharp rail contour is established,and a mathematical model for measuring the surface difference,gap and rail head width between the sharp rail and the basic rail is constructed,and the fusion algorithm of Zhang-Suen and gray center of gravity method is used to extract and refine the laser stripe,and finally obtains the center line of the sub-pixel laser stripes.In this experiment,the measurement was carried out on the straight sharp rail of the 18 th turnout with a speed of 250 km/h.The data obtained is basically consistent with the data of the contact measurement,which verifies the effectiveness and correctness of the method,and realizes the non-contact detection of the outline size of the sharp rail.

关 键 词：机器视觉 尖轨 面差 间隙 轨头宽度 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]