Research On Normal Consitency Of Point Cloud And Remeshing Based On Particle Opimization

3D digital modeling technology is an important means to obtain 3D shape of objects for a computer.It has been widely used in many fields such as computer animation design,virtual reality,finite element analysis,cultural relics protection and restoration.3D point cloud data is one of the most important input during 3D modeling.With the updating of scanning equipments and development of scanning technologies,it becomes more and more convenient and efficient to get 3D point cloud data.However,due to the accuracy of the scanning device,the illumination condition during the scanning process,the surface material of the object,and the self-occlusion of the object,the acquired 3D point cloud usually has problems such as uneven distribution,noise and data loss.The surface normal of the object is a very important factor in digital geometric modeling and digital geometry processing.For 3D point could expressed model,how to calculate the global consistent normal from the point cloud has always been a fundamental problem to be solved in modeling process.Especially when the point cloud is not closed and the model contains thin plates,how to deal with the global consistency of the normal is still a challenging problem.In addition,the 3D mesh obtained by modeling technology generally has the problems of large number of basic elements and poor mesh quality.It is difficult to directly apply to fields that are highly dependent on mesh quality,such as finite element analysis and texture mapping.How to re-mesh the 3D mesh quickly and with high quality becomes another important problem that needs to be solved in 3D modeling technology.Aiming at the above two basic problems,this dissertation pays attention to the global normal orientation problem on point cloud with missing data,global normal orientation problem on point cloud with noise and thin structure,and efficient remeshing algorithm on the triangular mesh model,based on the particle distribution optimization.The main contributions are as follows:(1)A novel space partition structure and its reconstruction method haven been proposed for orientating normal on non-closed model.We propose a novel space partition structure reconstruction method,named iBOT,in order to handle with scanning models reconstructed from significant missing point cloud.This method overcomes problems that BOT cannot deal with non-closed point cloud.(2)An algorithm on normal consistency for point cloud model has been pro-posed using particle distribution optimization.We propose a novel method for equidistant surface based on particle distri-bution.which can be used to calculate the equidistant surface for noisy or thin structured point model without reconstructing explicit representation.With the help of equidistant surface represented in the form of points,consistent normal orientation can be obtained,which improves the orientation accuracy on point cloud with thin structure.(3)An efficient remeshing algorithm has been proposed based on the power Voronoi diagram.A method for controlling the movement of particles on the tangent plane of the model is proposed.The weighted Voronoi diagram id constructed by using the matched mass points of different masses to avoid the computation of intersection between models surface and Voronoi cells.The algorithm execution efficiency based on the center of gravity Voronoi partitioning framework is improved,while ensuring the mesh quality.