Research On Surface Point Cloud Modeling And Accurate Extraction Method Of Deformation In Mining Area

Three-dimensional laser scanning technology is an important technology in deformation monitoring applications.It is an important research direction to analyze the surface deformation and the stability of structures from the acquired point cloud data of the ground and structures.Aiming at the structure and structure point cloud data,the mechanism feature point extraction algorithm is complex,the degree of automation is low,and the feature information is easy to lose;the point cloud filtering algorithm has low self-adaptation,poor practicability and low filtering accuracy;the ground deformation information extraction algorithm has poor reliability and poor data utilization.Sufficient and other issues,proposed the corresponding point cloud data processing algorithm,and combined with traditional measurement methods to verify the reliability and practicability of the algorithm,and achieved certain research results.The main research contents and results of this article include:(1)Research on the feature information extraction algorithm and deformation information calculation of building structures.Proposes an algorithm for 3D feature extraction of building point cloud based on Morse-Smale theory,which realizes fine,self-adaptive and highly automated extraction of feature points and feature lines of building structures;realizes 3D feature surface points by improving the region growth algorithm Fast extraction of cloud data;by fitting the characteristic straight line,calculating the slope and then calculating the tilt of the building,it is proposed to use the position change of the characteristic point to obtain the horizontal displacement and the sinking value of the building;it is realized on the basis of the characteristic line extraction algorithm The joint drawing with AutoCAD realizes the automatic drawing of the three-dimensional contour of the object.(2)Research the point cloud filtering algorithm.Based on the different spatial distribution characteristics of point cloud data in different dimensions,a linear discriminant classifier is constructed to classify vegetation point cloud and ground point cloud data,and a surface point cloud data filtering algorithm based on point cloud spatial distribution characteristics is realized.Compared with the existing filtering algorithm,the results of this algorithm show that it can achieve less parameter setting,high degree of adaptation,excellent filtering effect,and strong practical ground point cloud data extraction under the condition of only the three-dimensional coordinate value of the point cloud.(3)Research on the deformation extraction algorithm of surface point cloud data.After analyzing the main algorithm model based on the surface point cloud computing deformation information and the three factors that affect the accuracy of the point cloud deformation value calculation,a multi-scale model-to-model distance point cloud deformation calculation algorithm(M2M algorithm)is proposed.The distance between the center points of the core point set projected on the normal vector is taken as the point cloud deformation distance.It effectively overcomes the influence of point cloud measurement error,point cloud roughness,and registration error on the calculation of deformation value,and a reliable terrain change estimation is obtained.In addition,the vertical M2M distance and multi-scale M2M distance are calculated from the surface point cloud data of the selected mining area,and the calculated results are compared with the measured data to obtain the accuracy of the algorithm for calculating the surface deformation information.Realize the calculation and expression of deformation information based on point cloud data at three levels of core point,section line,and regional surface.Figure[89]table[10]reference[83]...