Interpolation Methods Of Large Scale Scattered Data

This paper presents the background and utility of Computational Visualization, andthe advances of research in mesh generation for finite element computation in last tenyears. The advanced of the common methods such as mapping methods, grid-basedmethods (include finite quadtree/octree methods), Delaunay trangulation and AFTmethods are emphasized. The main research fields such as surface meshing (include),all-hex meshing and parallel mesh generation are discussed in detail.This paper presents and discusses the Delaunay triangulation methods in details, suchas the definition and the characters of the Voronio picture and the Delaunay methods,and some useful constructed methods of Delaunay triangulation methods, such asA Delaunay triangulation method is brought forward oriented massive data, whichbased on the grid patition method. It divides the data set into some grid tiles,constructs Delaunay triangulation for each grid tile by divide-and-conquer algorithmbased on self-adapt grid partition, and store some unaffected triangles, then mergesadjacent Delaunay triangulations to whole or whole-like Delaunay triangulation. Thismethod requires low computer hardware, fits for parallel processing, can processDelaunay triangulation of massive data. At last, the trends of mesh generation arepresented briefly.