Metro Infrastructure Detection Based On Laser Scanning Imaging And Point Cloud Processing

Service status detection of metro infrastructure is a significant means to ensure long-term safety and stable operation of metro.With the rapid growth of metro operation system,traditional manual detection can no longer meet the current demand due to long detection time and human influence on detection results.Detection technology based on 3D machine vision has become an important means of metro infrastructure detection with its advantages of high precision,high efficiency and non-contact.Therefore,this thesis combined with the development project of rail inspection vehicle of a company under CRRC group,the detection method of metro infrastructure based on laser scanning imaging and point cloud processing is studied.The main research contents of this thesis are as follows:(1)A three-dimensional line structured light imaging component for metro infrastructure is designed.According to the metro operating environment,the principle of linear structured light measurement is adopted,and the suitable industrial camera,lens and laser are selected to form a three-dimensional line structured light imaging component.By using four line structured light imaging components,the field of view coverage of metro track area,evacuation platform area and contact rail area is realized.(2)A three-dimensional imaging method of metro infrastructure based on multi-view line structured light is studied.Firstly,aiming at the difficulty of calibration caused by lens distortion in this thesis,this thesis proved by derivation that the lens distortion link of the traditional nonlinear camera model can be post-positioned when the camera internal parameter matrix is a shape preserving matrix.Then,the camera lens distortion is obtained by using the straight line optimization method,in addition,the other camera parameters and light plane parameters are calibrated by using the sawtooth target method.Experiments show that after calibration by this method,the imaging accuracy of the linear structured light imaging component can reach 0.6mm,which meets the actual engineering requirements.Secondly,after completing the calibration of imaging components,a global calibration method based on portable CMM is studied.A portable CMM is used to establish a global coordinate system,a spherical target is used to provide global calibration feature points,and the transformation relationship from the local coordinate system of each component to the global coordinate system is obtained by least squares calculation,so as to unify the measurement data collected from different perspectives into the same coordinate system.Furthermore,the method of extracting the center of the light strip is studied.Combined with the actual situation,this thesis chooses to use the gray center of gravity method to extract the strip center.Finally,the influence of vibration on the point cloud collected by the component is studied.The point cloud of each component in the track cross section can be obtained by the two steps of coordinate system transformation and contour projection,and then combined with the mileage information,the point cloud splicing is realized.(3)After obtaining the 3D point cloud data of the infrastructure along the metro line,in order to meet the needs of the service state detection of the metro evacuation platform,the deformation degree detection algorithm of the evacuation platform and the overrun detection algorithm of the evacuation platform are studied.For the detection of the deformation degree of the evacuation platform,point cloud filtering is firstly performed on the collected three-dimensional point cloud.According to the placement method of the pedestrian walk board on the evacuation platform,the steel support positioning algorithm based on grayscale projection is used to realize the positioning of the steel support before and after the pedestrian walk board.Then,a three-dimensional point cloud of a single pedestrian walking board is obtained,and the edge detection algorithm of the point cloud based on the orthogonal plane constraint is used to realize the effective detection of the lower edge of the pedestrian walking board.Finally,the deformation degree of the evacuation platform is obtained.For overrun detection,a overrun detection algorithm based on line segment intersection is studied,which avoids the special intersection of traditional ray intersection algorithm and improves the operation speed of overrun detection.After experiments,the metro infrastructure detection system in this thesis has completed the high-speed laser scanning imaging of the infrastructure along the metro line,and the 3D point cloud data of the metro infrastructure in the track coordinate system is obtained.Compared with the manual detection results,the detection error of the deformation degree of the evacuation platform in this thesis is controlled within± 2mm.Compared with the traditional detection method,the time consumption of overrun detection is reduced by 27.44%.