Research On Surface Reconstruction Of 3D Laser Point Cloud

With the rapid development of 3D laser scanning technology,it has been widely used in deformation monitoring,intelligent city construction,cultural heritage protection and disaster assessment.The technology can quickly obtain the surface points of objects,and the original data mainly contains the coordinate values of points,which are massive and noise included.It must be processed before it can be used.There are a lot of noise points in the original point cloud,so it is necessary to denoise the original point cloud before the subsequent data processing.The large amount of the original point cloud makes the data processing time-consuming and needs to be simplified seriously.After the pre-processing is completed,it is necessary to study the surface reconstruction algorithm of the point cloud data in order to achieve the goal of surface reconstruction accurately and efficiently.Based on above analysis,the laser point cloud data processing algorithm is studied deeply in this paper in three aspects: point cloud denoising,point cloud simplification and surface reconstruction.Here are the main research contents and contributions:(1)Firstly,the data structure and geometric properties of point cloud are studied,the neighborhood partition,search of point cloud and the estimation of normal of point cloud are discussed and analyzed,and the advantages and disadvantages of traditional point cloud denoising algorithms are compared.Above all,based on the idea of surface fitting,a method of removing small-scale noise is proposed on the basis of the principal curvatures of point cloud to determine the type of fitting surface.The experiment shows that the method can effectively reduce the roughness of the point cloud surface and achieve the aim of smoothing of the point cloud surface.(2)In order to solve the problem that the point cloud simplification will easily cause the loss of quality while reducing the amount of point cloud data,a point cloud simplification method based on the difference of normal vectors in different neighborhoods is proposed in this paper.This method divides point cloud regions based on the difference of normal vectors of point cloud in different neighborhoods,and then simplifies point cloud data by the voxelized grid approach.Experiments show that the algorithm can not only effectively simplify the point cloud data,but also keep the features intact.(3)Combining point cloud denoising and point cloud simplification with mesh construction method of dimension-reduced Delaunay triangulation,a method of surface reconstruction based on Delaunay triangulation is realized,which removes small amplitude noise and simplifies point cloud.The results of reconstruction are compared and analyzed.Experiments show that this method can achieve better results.