Accurate Measurement Of Key Geometry Dimensions Of Railway Slab Using LiDAR Point Cloud

CRTS Ⅲ track slab is a new type of unit slab with shoulder,which has independent intellectual property rights in our country.Its integral forming and geometric dimension precision of finished slab have important influence on track smoothness.Current specification require that the geometric dimension processing deviation of railway slab be tested before leaving the factory.Based on intelligent robot and high-precision scanner,the automatic detection system of CRTS Ⅲ railway slab geometric dimension deviation for high-speed railway is used to detect the geometric dimension deviation of CRTS Ⅲ railway slab processing.The main idea is that to obtain the point clouds of each rail-bearing platform by the robot control scanner,determine the position and attitude of the point clouds,and then measure the geometric dimension of railway slab.Accurate determination of the position and attitude of the rail-bearing platform point clouds is one of the key issues to determine the precision of the system.Based on the three-dimensional surface registration technology,this paper studies this problem,and further improves the precision on the basis of the initial determination of the position and attitude of the rail-bearing platform point cloud by the robot.Based on the existing three-dimensional surface registration technology,a registration algorithm of bearing platform point clouds with additional constraints is proposed to solve the problem that accurately determining the position and attitude of rail-bearing platform point clouds obtained by blocks.By introducing the location of the embedded casing as a constraint,the algorithm solves the problem that the nearest point iteration(ICP)algorithm can not complete the registration due to the single longitudinal characteristics of the bearing platform surface.The simulation and experimental results using simulation data and measured data show that the RMSE of the proposed algorithm is about 0.05 mm after registration.According to the structural characteristics of different CRTS Ⅲ railway slabs,a unified model of CRTS Ⅲ track slabs based on sleepers and a coordinate system of railway slabs are proposed,which lays a foundation for the construction of a unified treatment method of CRTS Ⅲ track slabs adapted to various sub-models.A parameter extraction algorithm for embedded casing based on M-estimation is proposed.Reducing the influence that there may be noise points such as scum on the surface of embedded casing,the location precise of embedded casing can be obtained and the deviation value of embedded casing can be determined.The experimental results based on the simulated and measured data show that the position deviation of the extracted pre-embedded casing is not more than 0.01mm;the mean deviation of the center distance of the pre-embedded casing on the same bearing platform is 0.01 mm,and the maximum deviation is not more than 0.164mm;and the mean deviation of the pre-embedded casing skew deviation is 0.79 mm,and the maximum deviation is not more than 1.56 mm.A point cloud downsampling scheme considering the structural characteristics of the rail-bearing platform is proposed.According to the different functions of different areas on the surface of the rail-bearing platform in the proposed registration algorithm,the downsampling parameters of the rail-bearing platform different areas are defined,different downsampling parameters are adopted to ensure the accuracy and efficiency of the registration algorithm.Through the experiment of P5600 track plate,the ratio of the bearing surface of the bearing platform to the two clamp surfaces is 1/64.After the point cloud of each bearing platform is registered by the registration method with additional constraints proposed in this paper,the measurement items of key geometric dimensions of track plate are divided into three categories: bearing platform,sleeper and railway slab.The measured data of three different sub-models of CRTSⅢ track slab,P5600,P4925 and P4856,are used to test the results.The results show that the measurement accuracy of key geometric dimensions of track slab of different sub-models is equal,the measurement precision of skew value is about 0.63 mm,and the measurement precision of other items is 0.15 mm.The average time of P5600 track slab measurement is about 153.74 s seconds;the maximum deviation between multiple measurements of the same track slab is 0.42 mm,and the absolute deviation between the other measurements is not more than 0.1mm.