Research on the line point cloud processing method for railway wheel profile with a laser profile sensor

Abstract

Wheels are crucial components of railway trains, measuring their parameters is extremely significant to the routine maintenance of train operation. In this study, a line point cloud data processing method is suggested for measuring the wheel profile with a laser profile sensor. Aiming at two common types of noise point in line point cloud data, the steps of the method are divided into two parts: 1)Establishing a discriminant function based on the Angle-Distance algorithm to remove defective points; 2)Applying a segmental filter to process the noise data: in the small radius of curvature segment of the wheel profile curve, using improved SG filter which takes into account the distance-based Gaussian weight (DGSG), and in the large radius of curvature segment and linear segment are processed by Gaussian filter and median filter respectively. For lab and in-service wheels, the measuring procedures are both used. The geometric parameters are measured from the processed data, which are then compared with standard data to verify the accuracy of measurements. The comparable experiment results show that the geometric parameters measured from processed data have higher accuracy than the original data.