Research On Three-dimensional Point Cloud Data Meshing Technology

With the development and progress of 3D digital technology,the application of 3D point cloud technology has become more extensive.Point cloud meshing,as an important part of 3D reconstruction technology,is to establish spatial topology on the basis of known point cloud data to express the surface information of the object model.Due to the existence of measurement blind areas and other reasons,the collected 3D point cloud model has data missing problems,so it is necessary to repair the holes in the model and restore the original information of the object surface.In this paper,point cloud meshing and hole repair techniques are investigated.Aiming at the problems of difficult to control the mesh growth range when reconstructing non-uniform point clouds with density,the existence of holes and local mesh misalignment in the reconstructed surface,a method of meshing non-uniform point cloud data based on adaptive neighborhood is studied.Firstly,the seed triangle is constructed based on the normal vector and distance information of the point cloud;secondly,an edge is randomly selected from the seed triangle,and the adaptive search sphere radius is set according to the point cloud density,and the neighborhood of two vertices on the edge is queried as the candidate point set to limit the initial mesh growth range;then,the normal vector and projection plane of the points in the edge are calculated based on the peripheral information,and the vertices and their neighborhoods are projected onto the two-dimensional plane;finally,the non-uniform point cloud data are projected onto the two-dimensional plane by the plane constraint criterion;Finally,the unqualified candidate points are eliminated by the planar constraint criterion and mapped back to 3D space to select the candidate points with the largest dihedral angle for growth by the spatial constraint criterion to gradually construct a complete mesh surface.The experimental results show that the algorithm can effectively control the mesh growth range and construct a mesh surface that expresses detailed features.For the problem of missing data in the point cloud model,a point cloud hole repair method based on normal vector and curvature is studied.Firstly,the hole boundary is identified based on the mesh surface,the direction of the hole boundary is calculated,and on this basis,the concavity of the boundary vertices is judged to obtain the hole boundary angle;secondly,the optimal boundary point is selected by considering the angle and curvature;then,the coordinates of the new vertices are calculated by using the point cloud normal vector and the distance of the adjacent boundary,and the new triangles are checked for reasonableness to realize the hole repair;finally,the hole repair is adjusted according to the Delaunay Finally,the new mesh surface is adjusted according to Delaunay’s criterion and the principle of mesh smoothing.The experimental results show that the algorithm is effective in repairing the hole area and can achieve a smooth transition between the repaired mesh surface and the surrounding mesh,and accurately recover the information of the original model.