Fast Topology Reconstruction Of Triangular Mesh Models And Multi-resolution Display Technology

With the continuous development of computer technology,mesh models are gradually used in various fields,including film and television animation,geographic map information,virtual reality,computer vision,computer graphics,computer aided design,finite element methods,computational geometry and so on.In engineering,the triangular mesh file with STL format is widely used to the transmission of triangular mesh models.With the improvement of model accuracy,the mesh file is gradually increasing the amount of data.As a result,a large amount of triangular mesh data does not only bring the huge challenge to the computer storage,calculation,and transmission,but also to a certain extent affect the display efficiency of the model,which results in the subsequent processing difficulty of the model.In the light of complex triangular mesh models,the research work presented in this paper is mainly divided into two parts: one is to improve the efficiency of reading STL files and mesh topology reconstruction;the other is to simplify the model through mesh simplification algorithm and to construct the multi-resolution representation of the model.In the aspect of reading file data and reconstruction,this paper proposes a method of multithread memory mapping.Use memory mapping technology to establish the mapping relationship between files and memory,which can reduce system I/O operations,and improve the speed of reading files.When the data quantity of a file is very big,this paper adopts the idea of subsection mapping,and reads the data of each section using multiple threads in a parallel method,at the same time to reconstruct the topological adjacency relations between triangular mesh elements.The experimental results show that the method in this paper can effectively improve the efficiency of reading file and topology reconstruction.In the aspect of multi-resolution model display,mesh simplification algorithm is used to obtain simplified models with different resolutions.Among many mesh simplification algorithms,the quadratic error measurement algorithm is simple in calculation and fast in simplification,which can maintain the overall characteristics of the model to a certain extent.However,when the amount of data in the model is relatively large,the simplification efficiency of the algorithm begins to decline with the increase of the amount of calculation,and the overall simplification effect of the model is relatively uniform.In this way,the local detail features of the model are easily lost when the simplification rate is large.Aiming at this problem,this paper applies regional segmentation technologies into mesh model simplification.First the region growing method is used to segment mesh surface before simplification.The triangle of each segmented region has similar geometric features after segmentation,which can make the second error calculation be conducted by each region area.As a result,the amount of calculation can be greatly reduced and the simplified speed of models has been effectively improved.As there is the problem of easily losing local features of models using quadratic error metrics algorithm,this paper proposes a simplified method of feature preservation.On the basis of quadratic error metrics,folded edge curvature and the length are considered.In the role of the total error of folding,to simplify the flat mesh region is priority,and to delay to simplify the triangles in the obvious feature area.Experimental results show that the mesh simplification algorithm proposed in this paper is better than the quadratic error measurement algorithm that commonly used in terms of simplification rate and simplification effects.