基于密度的K邻近空间球算法研究  

Research of Spatial Sphere Algorithm for Searching K-Nearest Neighbors Based on Density

作　　者：赵灿[1] 傅丽[1] 程俊廷[1] 

机构地区：[1]黑龙江科技学院机械工程学院,哈尔滨150027

出　　处：《组合机床与自动化加工技术》2008年第12期25-27,31,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：黑龙江省自然科学基金重点项目(ZJG0607-01);黑龙江省普通高等学校青年学术骨干支持计划项目(1152G036)

摘　　要：提出了一种基于密度的K邻近空间球搜索算法。这种算法先根据空间数据的范围和数据点的总数对点云进行空间划分,再以当前被测点作为球心建立一个虚拟空间球,球体半径直接用K邻近点密度确定,并在与该球干涉的立方体栅格中搜索K邻域。实验证明,所提出的算法有效地提高了K邻近搜索的效率,对不同拓扑结构和不同疏密程度的点云具有稳定性。A spatial sphere algorithm for searching K-Nearest Neighbors based on density is presented. At first,the scattered points are divided into a set of uniform cell by considering the range of data set and the total numbers. Then suppose a dummy spatial sphere with the center as the current point, and the radius decided by the density of K-nearest points directly. Search the K-Nearest Neighbors in the cube grid interfered by the sphere. Experiments show that the algorithm effectively enhanced the efficiency of searching K-Nearest Neighbors, and has stability for the point cloud with the different structure and the different density.

关 键 词：逆向工程 K邻近 点云 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]