三维曲面部分匹配的算法研究

将基于曲率分析的曲面片形状划分方法和几何哈希相结合，提出一种通用的空间曲面匹配算法。对待匹配的三维曲面模型，计算其网格顶点的主曲率和法矢；由主曲率和该点所在的曲面片类型来构造其无向脚标，有向脚标为该点的法矢。按照多重筛选标准生成数量少却有效的匹配点对，建立候选点对列表。由候选点对所生成的三维空间变换组成哈希表，运用双层哈希投票机制得出使模型能够正确匹配的三维坐标变换矩阵。实验表明，该算法适用于具有部分重叠的曲面模型的匹配，并能保证较好的匹配精度和速度。

Algorithm for Three-Dimensional Partial Surface Matching

A uniform algorithm for partial surface matching is prop osed. It combines geometric Hashing scheme with surface segmentation based on cu rvature analysis. Given two objects for matching three dimensional surface, the principal curvatures and the normal vector at each mesh vertex are calculated. T he undirected footprint consists of principal curvatures and its patch type. Nor mal vector is the directed foot-print. A multi-filtering criterion ensures the matching point-pair to be compact and effective, meanwhile the list of candida te points is constructed. A hash table is constituted from the coordinate transf ormations in three dimensions corresponding to these candidate points. A two-le vel voting scheme is employed to find out the target transformation that makes a sufficiently large portion of one object close to corresponding portion of the second object. Experimental results show that the algorithm is suitable for par tial surface matching. So it also is accurate and efficient.