摘要
针对基于传统径向基函数的数据插值方法在重建大量数据点云曲面时的困难,提出将数据点先分割再分别重建的方法.将点云的包围盒沿坐标轴分割,两两合并相邻的方块,使得方块相互重叠且覆盖整个包围盒.对每个包围盒内的点用径向基函数方法插值,利用窗口函数将每个函数限制在各小方盒内求和得到最终的整体插值曲面.借助MC(marching cubes)方法得到三角网格曲面.每个方块内点云的重建过程可以并行实现,因此该方法非常适用于对重建效率要求较高的场合.
Radial basis functions based data interpolation was introduced.Aiming at the problems of reconstruct surfaces from large data points,a method which partitioned point clouds and reconstructed separately was presented.After the bounding box of the point cloud is partitioned along axes,every two neighbors are united,such that the obtained boxes cover the whole bounding box.Radial basis functions are interpolated for the points in every small box,and an implicit function is obtained by summing all the functions restricted by window functions.Simplicial mesh is obtained by marching cubes.The method is very suitable to be applied to fast reconstruction system because it can reconstruct surfaces in small boxes in parallel.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2010年第4期728-731,共4页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(60773179)
国家"973"重点基础研究发展规划资助项目(2004CB318000)
关键词
点云
曲面重建
径向基函数(RBF)
分割
point cloud
surface reconstruction
radial basis function （RBF）
partition
