The Optimization of Triangulation based upon Serial Fault Data
CT fault data is commonly used in surface reconstruction in the field of reverse engineering. The three-dimensional reconstruction based upon serial fault data is divided into two parts, namely boundary contour splice and end contour closure. In the part of boundary contour splice, Delaunay triangulation is the common algorithm. However, in practical application, when large difference of data number or contour distance appears between two adjacent fault data, Delaunay triangulation algorithm is prone to generate long-narrow triangles or radial family. Thus, this paper puts forward Delaunay triangulation algorithm based upon shortest distance first principle. In the part of end contour closure, Delaunay triangulation algorithm based upon determination of convex-concave vertices is more commonly used. End contour closure actually means triangulating in the polygon, and in computer graphics, the triangulation of any polygon contour will affect the quality of three-dimensional reconstruction to a certain extent. The end contour from CT image is generally comparatively complex, and its data points are very dense. While Delaunay triangulation based upon determination of convex-concave vertices is applied to CT image, it is prone to generate long-narrow triangles and triangles whose sizes differ greatly, even in some cases it cant be triangulated completely.Therefore, this paper puts forward an improved algorithm which combines Delaunay triangulation based upon determination of convex-concave vertices with interpolation. Research shows that the improved algorithms not only maintain the original advantages but also further raise the effectiveness of triangulation, and enhance the universality of algorithms.
CT data Delaunay triangulation Surface reconstruction Boundary contour End contour
Yang Xiaodong Liu Peng
Dept.of Mechanical Engineering and Automation,Harbin Institute of Technology, Harbin,150001
国际会议
北京
英文
261-266
2008-11-06（万方平台首次上网日期，不代表论文的发表时间）