3D Deformation Monitoring And Data Processing Of Mining Area Surface Based On UAV-LiDAR

UAV-LiDAR is a remote sensing technology with strong autonomy,which can monitor target changes in a timely and effective manner.Using it to monitor surface movement in mining areas has the advantages of accuracy and efficiency.The UAVLiDAR point cloud is often used to construct the DEM model of the mining area,and the DEM difference method is used to obtain the surface deformation in the vertical direction.However,the ground surface movement in the mining area is a complex three-dimensional deformation process.It is difficult to obtain the surface deformation in the vertical direction to meet the diversified monitoring needs of the subsequent mining area,and the plane position information of the UAV-LiDAR point cloud cannot be used.It is urgent to find an effective method to utilize all the threedimensional spatial information of the point cloud and obtain the three-dimensional mobile deformation field of the ground surface from the UAV-LiDAR point cloud.Therefore,based on UAV-LiDAR technology,this paper simultaneously acquires the three-dimensional movement of the mining area surface,uses the spatial characteristics of the point cloud to study and solve the problem of three-dimensional deformation extraction,and applies these research results to the actual monitoring of coal mine surface deformation.The focus is on the data processing of UAV-LiDAR point cloud,and the applicability of different methods to extract the three-dimensional movement deformation of the mining area surface is discussed.The main achievements are as follows:(1)Aiming at the immature problem of UAV-LiDAR monitoring technology in mining area surface monitoring,starting from the principle of UAV-LiDAR monitoring technology,a complete set of UAV-LiDAR monitoring mining area deformation scheme and data processing flow is proposed.At the same time,according to the error source and calculation of its measurement accuracy,the measurement accuracy is between 18 mm and 42 mm,which verifies the feasibility of this technology applied to the monitoring of surface deformation in mining areas.(2)Taking four kinds of topographic point clouds as experimental data,three filtering algorithms of PTD,CSF and SMRF are used to obtain surface points,and three interpolation algorithms of IDW,Kriging and TLI are used to construct DEM.In order to avoid Type II errors and obtain high-precision DEM efficiently,the combination of PTD filtering and IDW interpolation is selected to obtain DEM.In view of the noise problem in the obtained subsidence basin,an improved non-local filtering denoising algorithm is adopted to filter the noise points as much as possible while retaining the original details of the basin.(3)For conventional monitoring,subsidence and horizontal movement are extracted separately.This method cannot reflect the synchronization of threedimensional surface movement while bringing errors.In this thesis,a rigid body registration algorithm based on connected domain segmentation is proposed to obtain the three-dimensional movement of the ground surface by using the characteristics of the rigid body in the mining area.The connected domain segmentation algorithm is used to segment the surface rigid body,and then the rigid body is combined with the FPFH feature and the Trimmed-ICP registration algorithm,and the M2C3 algorithm is used to correct the error,and the three-dimensional movement of the ground surface in the experimental area is extracted.(4)The CPD non-rigid registration algorithm based on direction constraints is introduced to register feature points,which is used to solve the problem that the rigid body in the mining area is not enough to obtain the three-dimensional deformation of the ground surface.Combined with the mobile registration unit technology,the whole is transformed into an individual unit for unit registration.Experiments show that the internal coincidence accuracy of the three-dimensional surface deformation X,Y,and Z obtained by the algorithm is 15 mm,17mm,and 27 mm,respectively,which is basically consistent with the laser radar measurement accuracy.(5)In the actual coal mine surface deformation monitoring,the threedimensional deformation results of the mining area are analyzed and the accuracy is evaluated.The experimental results show that the displacement accuracy of the method proposed in this thesis is 61 mm,41 mm and 24 mm respectively in the eastwest,north-south and vertical directions.