With the rapid development of high speed railway and the increased transportation load,the requirements for the reliability of railway infrastructure become higher and higher.In order to realize comprehensive surface defect detection for railway in real time,a rail surface defect detection system based on binocular structure light is proposed.The system mainly involves the construction of 3D data acquisition equipment and the realization of rail surface defect detection algorithm.In the building process of data acquisition equipment,a binocular system is composed with two adjacent cameras,the line laser is projected by laser launcher and camera collects the images of the shape change of laser. In the process of reconstruction,adjacent internal and external parameters are calibrated and obtained,and in accordance with the Gaussian fitting and the pole line constraint methods, the whole surface of rail is reconstructed in three dimensions.Through the simulation of rail movement,the reconstruction of rail sample is completed.Compared with reconstruction data and effects,it is found that this method can extract the information of depth of the rail surface defects effectively,and the measurement error is limited within 4%,which satisfies the accuracy requirements for rail surface defect detection.%随着高速铁路的快速发展及运输负载的增加,国家对铁路基础设施可靠性的要求越来越高.为实现钢轨表面缺陷的实时?多角度探测,设计了基于双目结构光的钢轨表面缺陷检测系统.该系统主要涉及三维数据采集设备的搭建及钢轨表面缺陷检测算法的实现.在搭建数据采集设备的过程中,由相邻的2 台摄像机组成双目系统,利用激光发射器向钢轨表面投射线激光,并由相机采集钢轨表面线激光的形状变化图像.在重建过程中,标定并获得相邻的内外参数,采用高斯拟合和极线约束等方法,实现了完整钢轨表面的三维重建.通过设计模拟钢轨运动,完成了钢轨样品的重建.对比重建数据及效果可知,该系统可以有效地提取钢轨表面缺陷的深度及轮廓信息,且测量误差为4%,符合钢轨表面缺陷检测的精度要求.
展开▼
