Research On Data Processing Method Of Tunnel Deformation Point Cloud Based On 3D Laser Scanning

Deformation measurement is of great significance to the structural safety of tunnel engineering.Traditional deformation measurement methods often rely on manpower and are mainly single-point and contact measurement methods.Therefore,it is difficult to respond to the needs of the new era in terms of efficiency and informatization.The high precision,high efficiency,high digitization,initiative,non-contactation and other characteristics of 3D laser scanning technology make it have the advantages that traditional deformation measurement methods can hardly match.In recent years,the application of 3D laser scanning technology in the field of tunnel engineering has continuously emerged,but there are still many shortcomings,and many problems need to be further explored.In this thesis,the point cloud data derived from the 3D laser scanner is used as the core data carrier of deformation measurement,and the problems of 3D laser scanning technology in point cloud data processing algorithm and tunnel deformation measurement application are discussed.A tunnel deformation measurement method based on point cloud feature matching is proposed in the dissertation.The registration algorithm is used to optimize the coordinate system normalization results derived from the 3D laser scanner,and the local point cloud model registration and normal mapping are combined to establish the relationship between corresponding point pairs.It has better measurement accuracy while retaining the "volume measurement" characteristics of point cloud data.First of all,for the data reduction problem of the tunnel point cloud model,this dissertation discusses and researches the downsampling algorithm and the point cloud denoising method respectively.Simulation experiments are used to compare the effects of different downsampling algorithms applied to the tunnel point cloud model,analyze the sampling efficiency and performance of each algorithm,and optimize the algorithm through the octree index.Combining statistical denoising and radius denoising algorithms,a distance denoising method based on iterative filtering model is proposed.This method is suitable for the tunnel point cloud model during the construction period with complex environment,and can effectively eliminate most of the point cloud noise.Secondly,research on the key technology of tunnel point cloud model registration method.The registration results of different global registration algorithms applied to the tunnel point cloud model are analyzed through registration experiments.Experiments show that the FGR algorithm is superior to the classic RANSAC algorithm in terms of registration accuracy,algorithm time-consuming,and robustness.According to the results of simulation experiments,the influence of the registration parameters on the point cloud registration effect is analyzed,and the key parameter value recommendations suitable for the tunnel point cloud model are given.Thirdly,in view of the shortcomings of the application of 3D laser scanning technology in the field of tunnel deformation measurement,this dissertation puts forward a local deformation measurement method of tunnel based on the research of point cloud registration as the core.The "FGR-ICP" registration strategy is used to optimize the registration results derived from the coordinate system normalization method,and then use the local point cloud model registration combined with the centroid normal line estimation to establish the corresponding relationship of the deformation points with the same name,and realize the tunnel deformation measurement.Compared with the general three-dimensional laser scanning deformation measurement method,this method optimizes the registration error generated by the coordinate system normalization,highlights the spatial threedimensionality and data comprehensiveness of the point cloud data,and does not depend on the measurement point layout.Finally,the reliability of the deformation measurement method proposed in this dissertation is verified by the multi-period scanning data during the construction of a certain tunnel.The results show that the average error of the deformation measurement method proposed in this dissertation and the field measured data is-2.46 mm,and the minimum error can reach-0.21 mm.Compared with the average measurement error of the coordinate system normalization method of-22.33 mm,this method has nearly an order of magnitude improvement in measurement accuracy.