Surface And Parameterization Of Automobile Seat Point Data

Abstracting with economic globalization, new technology revolution and industry restructuring, the internationalization trend in automobile production has become more and more distinct. As the direct supporting device of automobile users, the seat is an extremely important auto interior component. Currently, joint ventures take the leading place in domestic auto seat market. Although the export of domestic auto seats is growing, the technology gap between domestic and oversea manufacturers is still a large one.This essay researches current research situation and developing trend of this technology, and points out that reverse engineering of objects is the key content of contemporary reverse engineering, and the key techniques of reverse engineering of objects is the most important thing. With the development of global economy, the consumption of auto products is increasing, and the demand for auto seats will also increase. The appearance of surface curving and parameterization softwares complies with such a trend, conforms to the law of auto development, and meets the needs of auto development and consumption market.The surface-curving and parameterization of auto seat data is the bottleneck of auto seat production. The main technology involved is Surface Modeling Technology, which requires advanced computer hardware and good ability of surface model design, and mainly consists of feature extracting technology of three-dimensional point cloud data and reconstructing technology of NURBS surface.This paper aims to realize NURBS surface reconstruction, and the automation and parameterization of auto seat design.First, this essay elaborates on acquisition and simplification of point cloud data, and compares various acquisition technologies and point cloud simplification algorithms and analyzes their advantages and disadvantages. Chord and angle-deviation criterion is simple but time-consuming and requires complex calculation, for a triangle has to be formed before relevant calculative processing is done; uniform grid model is very useful in processing point cloud comparison data by calculating the distance between every point to center-point to find k-points inside the uniform grids, but it is limited in processing complex model data and does not meet the practical demand of the software; Lisen information content model captures much information of point cloud data and requires complicated calculation; finally, based on existed algorithms, this paper proposes its own simplification algorithm of three-dimensional point cloud data and reaches its aim of cutting the amount of calculation and improving data processing efficiency.This paper adopts the algorithm for transition from point-cloud data to triangular mesh. Triangulation of scattered point-cloud data on spatial surface is an important method for scientific calculation and analysis, which is widely used in the interpolating surface construction of scattered cloud data and the visualization of 3D data field. There are plenty of triangulation algorithms for point-cloud data processing, and most of them generate the Delaunay triangle mesh.Lawson Method is mainly based on three configurations of four two-dimensional points and two kinds of transformation between configurations, starting from an unoptimizable triangulation and making empty-circumcircle check on adjacent triangles sharing one side, and its triangulation mainly directs at two-dimensional point cloud. Basing its three-dimensional Delaunay triangulation algorithm on surface switching, Joe Method is not merely an extension of Lawson Method, but combines with point-adding method and constructs an edge-face-exchange algorithm with complexity. Joe Method is based on five configurations of five three-dimensional points and four kinds of transformation between configurations. Combined with point-adding method, Joe Method obtains high efficiency and uniformity, for all points to be dissected have become interior points of auxiliary tetrahedron. This paper adopts Joe Method.Region growing is to be done after triangle mesh is finished. This paper introduces how to extract the feature line and how to perform region-growing and finally realize surface reconstruction. The algorithm for feature line extraction from triangular meshes based on STL files, proposed by Yang Shengyuan, etc., is mainly based on dihedral angels of triangle mesh and realizes region parting by extracting feature line through dihedral angels. The algorithm of extracting feature line from triangular mesh model, proposed by Liu Shenglan,etc., is mainly based on the basic idea: the triangular mesh type is marked with curvature of triangle, and then the triangular meshes which have the same marker are merged, finally the triangular meshes are optimized continuously. The core of the approach to creating-surface by characteristics-line, proposed by Tang Jingxiong, etc., is to use angle between two vectors extract characteristics-lines; use curvature values marked triangular grid, merge the same mark triangular mesh, as encountering the characteristics-lines, stop the merging; carry out sub-block into four areas; finally use Coons surface get a series of NURBS patches.Through a comparison between the above-mentioned algorithms, this paper proposes a region growing algorithm based on dihedral angels to realize NURBS surface. The approach is: acquire the dihedral angel between adjacent triangles; then, compare dihedral angels of triangle mesh and see whether they correspond with the prescribe threshold value: if yes, combine the two triangle meshes, otherwise stop the growing and continue on another triangle mesh until all meshed finish region growing; Although the algorithm result in this article is the polygon region, because of handling specific point data of the car seat, this article makes full use of the actual features of the car seat: surfaces of the dihedral angle of the small range, and the adjacent surfaces of the dihedral angle of the big small, so, the polygon regions gotten by algorithm in this article is actual quadrilateral regions with the feature lines, while omitting to transform from the polygon region to the quadrilateral region transformation; this algorithm improves the efficiency; the corresponding control vertexes are acquired and NURBS surfaces come into being.At the stage of parameterization, this paper introduces a few methods and suggestions from Shi Fazhong on NURBS surface modification, and combing with practical requirements in the following practice, makes adjustments on control vertexes and control nodes of NURBS surface, and finally changes the shape of NURBS surface to meet working demands.This paper realizes the software through which 3D object models can be obtained after reading object point cloud data. It saves designers'effort in learning to use complicated CAD design softwares and thus lowers the cost and reduces design time and design cycle.