• CN: 11-2187/TH
  • ISSN: 0577-6686

Journal of Mechanical Engineering ›› 2025, Vol. 62 ›› Issue (6): 237-246.doi: 10.3901/JME.260190

Previous Articles    

Research on the Measurement Method of Value Reduced for Switch Rail Based on Line Laser Sensor

QIN Bin1, LIU Yueming1, DING Yixian1, HOU Luyao1, CHEN Guopeng2   

  1. 1. School of Mechanical and Electronic Control Engineering, Beijing Jiaotong University, Beijing 100044;
    2. CRRC Technology Innovation (Beijing) Co., Ltd., Beijing 100010
  • Received:2025-05-04 Revised:2025-11-29 Published:2026-05-12

Abstract: Switch wear directly affects the stability and safety of trains passing through turnout. Value reduced for switch rail is a key indicator of switch wear, and its measurement accuracy is a crucial prerequisite for ensuring the normal operation and safety of switch rail. To solve the problem of low detection accuracy and difficulty in ensuring repeatability in the current measurement of value reduced for switch rail, a method based on line laser sensors is proposed. First, point cloud of the switch rail collected from two angles are processed. Redundant noise and edge noise are removed using pass-through filtering and radius filtering, respectively, while bilateral filtering is used to smooth the point cloud. Point cloud simplification is achieved using farthest point sampling and random sampling, depending on the curvature of the points. Next, the common regions from two acquisition angles are extracted, and the point clouds are coarsely aligned using the PCA algorithm. Fine alignment is then achieved using a point-to-plane ICP algorithm optimized with a robust loss function. Finally, a NURBS curve is employed for curve fitting of the switch rail, enabling the calculation of the value reduced for switch rail. Comparative experiments validate the effectiveness of this measurement method, showing that the repeat measurement errors for 5mm and 20mm cross-sections are reduced by 42.6% and 43.3%, respectively.

Key words: turnout, value reduced for switch rail, line laser sensors, point cloud, curve reconstruction

CLC Number: