基于弦测法与密度聚类的三维结构光波磨检测

为解决钢轨波磨人工检测费时费力及惯性法检测精度较低的问题,综合利用三维结构光技术、弦测法和密度聚类算法进行波磨检测。首先获取钢轨点云,通过纵向平面遍历轨头点云得到钢轨的纵向截面簇,其次利用弦测法计算相应纵向截面的谷深和波长。最后利用钢轨发生波磨时产生的接触斑,对遍历计算的谷深和波长分别进行密度聚类,将相同或相似接触斑的波磨信息聚类成簇,对聚类后各簇的结果进行统计分析得到相应钢轨的波磨信息。本方法结合三维结构光数据量大、弦测法计算直接明确的优点,通过密度聚类将不同深度、大小的波磨接触斑进行区别,有效地将钢轨波磨不同谷深及波长成分进行分类计算。实验室样件试验及现场试验表明,本方法能够实现对钢轨波磨的精确检测。

Three-dimensional Structured Light Corrugation Detection Based on Chord Measuring Method and Density Clustering

In order to solve the problem of time-consuming and laborious manual detection as well as low detection accuracy of the inertial method in corrugation detection, this paper proposes the corrugation detection by comprehensive use of three-dimensional structured light technology, chord measuring method and density clustering algorithm. First, the point cloud of the rail is obtained, and the longitudinal section cluster of the rail is obtained by traversing the point cloud of the railhead of longitudinal plane. Second, the valley depth and the wavelength of the corresponding longitudinal section are calculated by chord measuring method. Finally, based on the contact patches generated during the corrugation, density clustering of the valley depth and the wavelength of the ergodic calculation are respectively conducted, and the wave information of the same or similar contact patch is clustered into clusters, and the corrugation information is obtained through statistical analysis of each cluster after clustering. This method combines the advantages of both the three-dimensional structured light with sufficient data and the chord measuring method of straightforward calculation, and distinguishes the corrugation contact patches of different depths and sizes by density clustering. Thus, different valley depths and wavelength components of corrugation can be effectively classified and calculated. Laboratory sample test and field test show that the method can accurately detect rail corrugation.