轨道列车车轮型面线点云数据处理方法研究

Research on Data Processing Method of Point Cloud of Rail Train Wheel Type Surface Line

随着轨道列车交通的迅速发展,为了保证列车运行的安全性,需要及时检测列车车轮型面的磨损。非接触式的激光检测装置由于其具有精度高、方便快捷等原因,逐渐应用到列车车轮检测的工作中。但在使用激光传感器扫描列车车轮型面时,由于车轮型面复杂导致光线的不均匀反射,传感器得到的车轮型面中存在部分异常点。异常点的存在会影响车轮型面参数的测量与车轮磨损程度判断,因此本文提出一种基于夹角-弓高-曲率联合判别的异常点方法,对车轮型面所有采样点建立统一偏差尺度,并采用移动窗口的方法对超出局部阈值的噪声点进行识别并剔除,再对剔除后的数据进行多项式滤波处理。实验结果表明,所提出的数据处理算法能够很好地剔除异常点,且处理后能够明显降低车轮型面参数的测量误差,方便车轮检修人员精准记录车轮数据,从而保障列车的安全运行。

With the rapid development of rail train traffic, in order to ensure the safety of train operation, it is necessary to detect the wear of train wheel profiles in time. The non-contact laser detection device is gradually applied to the work of train wheel inspection due to its high precision, convenience and speed. However, when scanning the wheel profile of the train with a laser sensor, there are some abnormal points in the wheel profile obtained by the sensor due to the uneven reflection of light due to the complex wheel profile. Therefore, this paper proposes an anomaly point method based on joint discrimination of angle-bow height-curvature, establishes a unified deviation scale for all sampling points of the wheel profile, and uses the method of moving window to identify and reject the noise points beyond the local threshold, and then performs polynomial filtering processing on the rejected data. The experimental results show that the proposed data processing algorithm can eliminate abnormal points well, and the measurement error of wheel profile parameters can be significantly reduced after processing, which is convenient for wheel maintenance personnel to accurately record wheel data, so as to ensure the safe operation of trains.