Research On Point Cloud Surface Reconstruction And Hole Repair Technology

The surface reconstruction of point cloud data is to restore the geometric topology of spatial points based on the existing point cloud data and restore the surface shape of the object model.It is an indispensable link in reverse engineering.During the process of collecting point cloud data using a laser scanner,due to model incompleteness,blind spots,etc.,local data will be missing,resulting in incomplete surface and holes in the generated 3D model.The point cloud hole repair is based on the data information of points near the hole,and fills the hole by generating new points through the filling algorithm to restore the original surface shape of the object model.This paper studies the technologies related to point cloud reconstruction and hole repair.In order to improve the accuracy and efficiency of point cloud surface reconstruction,a scattered point cloud surface reconstruction method combining plane projection and area growth is proposed.Firstly,a local point set is selected from the scattered point cloud and its dispersion is performed Judge,project a smoother set of points that meet the requirements of dispersion onto a two-dimensional plane and perform triangulation,and map the topological connection relationship between the points after triangulation back to the three-dimensional space to realize the surface reconstruction of the point cloud.Then identify The boundary of the generated mesh is developed,and the remaining scattered points are grown outward along the boundary with an improved area growth method until all the scattered points are incorporated into the triangular grid.This algorithm can reconstruct the structural form A 3D model that is accurate and retains the details of the object,reduces the complexity of surface reconstruction and improves its accuracy.In order to improve the surface integrity of the 3D point cloud model,a point cloud hole repair method combining an ordered edge table method and a radial basis function is proposed.Boundary points are first extracted from the triangular mesh,the boundary points are projected onto a two-dimensional plane,and filled points are added to the hole using the ordered edge table method and the mesh is triangulated,and the coordinates of the filled points are mapped back to three-dimensional space.Then,extract the constraint points and determine the normal vector of the points,establish a radial basis function,adjust the added filling points layer by layer and smoothing with Laplace algorithm.This algorithm has a better filling effect for holes.The filling points are evenly distributed in space and connected smoothly to the original hole boundary,which can more accurately restore the morphological characteristics of the original model.The point cloud display platform was written in C++ under Win7 environment,used OpenGL graphics library to render point clouds and grids,used MFC framework to implement visual interface,which can display the algorithm results of this article.