| The spatial position data of each rail bearing platform of the CRTSⅢ track slab after the laying is completed is one of the important basic datas of track fine adjustment for High-Speed Railway.At present,the CRTSⅢ track slab is mainly carried out by manual on-site measurement,which has disadvantages such as heavy workload and low efficiency,and it has been difficult to meet the needs of my country’s High-Speed Railway construction.Terrestrial Laser Scan(TLS)technology has the advantages of fast scanning speed and high accuracy.It can quickly scan to obtain high-precision track slab point cloud,and then complete the track slab measurement through point cloud processing.As the track slab measurement must use the engineering coordinate system,and the point cloud acquired by each station is the local coordinate system,how to accurately determine the engineering coordinates of the CRTSⅢ track slab point cloud acquired is a key issue.The research work of this paper mainly includes the following two aspects:(1)Point cloud stitching of CRTSⅢ track slab based on plane feature constraintsDue to the limitation of scanning angle and effective scanning distance during TLS scanning,the CRTSⅢ track slab point cloud between two pairs of CPⅢ control points needs to be collected by multiple stations.The similarity of CRTSⅢ track slab point cloud makes it difficult to directly use the Iterative Closest Point method(ICP)for stitching.In addition,components such as the rail bearing platform on the track slab are easily affected by the scanning angle,which makes the point cloud of the same rail bearing platform scanned by adjacent stations have symmetry,so makes it difficult to realize the splicing of the CRTSⅢ track slab point cloud of the two stations through global or local features.Since the planes of the CRTSⅢ track slab point cloud are not affected by the scanning angle,and the point pairs in the planes with the same name in the overlapping area of the track slab point cloud of adjacent stations can accurately represent the spatial position of the two station point cloud.Based on this feature,the two-station CRTSⅢ track slab point cloud splicing method with plane feature constraints is proposed.This method first realizes the initial splicing manually,and then reconstructs the point pairs with the same name based on the same-named planes of the CRTSⅢ track slab point cloud of the two stations,and uses the weighted SVD conversion matrix solution method to calculate the conversion matrix to realize the precise splicing of the point cloud.The proposed method is tested based on two sets of measured CRTSⅢ track slab point cloud and compared with the Generalized Iterative Closest Point method(GICP).The experimental result shows that the proposed method can realize the splicing of CRTSⅢ track slab point cloud,and compared with the GICP method,its splicing accuracy is improved by16%.A pair of CPⅢ control points are arranged every 60 m along the High-Speed Railway,and three measuring stations are usually required to be set up between the two pairs of CPⅢ control points to complete the point cloud collection of the CRTSⅢ track slab.The two-station CRTSⅢ track slab point cloud splicing method with plane feature constraints is used to splice the front and back station point cloud to the middle station,the complete track slab point cloud between the two pairs of CPⅢ control points can be obtained,and this will lay a foundation for the determination of the engineering coordinates of the point cloud of the track slab.(2)Robust coordinate conversion of track slab point cloudAiming at the problem of there are possible gross errors in CPⅢ control points coordinates,the robust engineering coordinate conversion method of CRTSⅢ track slab point cloud is proposed to determine the engineering coordinates of the complete track slab point cloud between two pairs of CPⅢ control points.An automatic extraction method of spherical targets with multiple geometric information constraints is first proposed to accurately extract the coordinates of all spherical targets in the point cloud,which solves the difficulty for the neighbourhood distance mutation method which has the radial background requirements to accurately extract all the pherical targets,especially directional spherical targets erected on a tripod in the High-Speed Railway scenes.Then,the robust Gauss-Newton coordinate conversion method is used to convert the the complete track slab point cloud into the engineering coordinate system.In addition,a high-precision conversion model of engineering coordinates and mileage is realized in the presence of multiple projection belts.The simulation experiment results show that when the CPⅢ control points have gross errors,the robust engineering coordinate conversion method of CRTSⅢ track slab point cloud can obtain the correct results,and the deviation of the results is less than 0.12 mm.The coordinate conversion test of CRTSⅢ track slab point cloud is carried out through three point clouds which are scanned in the simulated track slab scene,and the spherical prisms are used as the accuracy checking points.In the experiment,the middle station is used as the reference station,and the point clouds of the front and back station are respectively spliced with it to obtain the spliced track slab point cloud.The result shows that the two-station CRTSⅢ track slab point cloud splicing method with plane feature constraints can effectively realize the splicing of CRTSⅢ track slab point cloud.When the robust engineering coordinate conversion method of CRTSⅢ track slab point cloud is used to determine the engineering coordinates of the spliced track slab point cloud,the accuracy of the coordinate conversion is calculated according to the coordinate difference of the spherical prisms.And compared with the Gauss-Newton coordinate conversion method,its accuracy of the coordinate conversion is improved by 7.7%.The experiment results show that the proposed method in this paper can accurately determine the engineering coordinates of the track slab point cloud. |
PDF Full Text Request