Research On Approaches To Extracting Feature Lines And Filling Holes For Triangular Mesh Models

With the rapid development of the technology of digital geometry and manufacturing, reverse engineering has been widely used in mechanical industry, art design and many other fields. Reverse engineering is a technology that the design principle of products can be deduced directly from the final product. As the triangular mesh models are the most commonly used in reverse engineering, extraction of feature lines and filling holes for triangular mesh models are the basis of reverse engineering and have important research significance.The triangle mesh model is the foundation of extraction of feature lines and filling holes. In this paper, in order to build the topology relationship of triangles in the STL format file, the triangle mesh model of STL format is reconstructed by using the data structure of coding half-edge and Balanced Binary Tree. A feature line extraction method combining mean curvature with mesh edges is presented for the mesh models of general mechanical parts. Firstly, the angle between the normal of adjacent triangles is used for the initial extraction of sharp edges. Then the feature edges are extracted according to the relationships between concave-convex properties and mean curvature of the surface. Finally the vertices of the extracted feature edges are classified based on corners and linked into feature lines. For the mesh models containing a lot of noise or the arts, this paper has used the higer order derivative of cuvature to extract the feature vertices. The first order derivative and second order derivative of curvature are the judgment conditions of the vertices. The holes leaved after segmentation of complex mesh models must be filled to describe geometric feature. The new triangles are created by using an Advancing Front method in the characteristic polygon firstly. Then a fitting surface based on the imformation of the vertices around the holes is used to adjust the new triangles. Finally the examination of the legitimacy and the optimization of meshes have been done.The algorithms in this paper have been achieved on the development platform of Visual C++2 008 and OpenGL. To verify the effectiveness of the algorithms, several examples are presented and compared.