基于有点区域的K近邻搜索新算法  

A new algorithm for searching K-nearest neighbors based on point area

作　　者：刘立强[1] 康宝生[1] 

机构地区：[1]西北大学信息科学与技术学院,陕西西安710127

出　　处：《西北大学学报（自然科学版）》2010年第5期784-787,共4页Journal of Northwest University（Natural Science Edition）

基　　金：陕西省自然科学基金资助项目(FC06121)

摘　　要：目的对不同的点云数据,根据有点区域占空间包围盒的比率的不同,对栅格划分最优边长进行估算。方法在分析已有栅格边长估算方法的基础上,提出基于有点区域的二次分割算法来估算最优边长。结果给出了算法的步骤,并将该算法应用于不同的点云数据进行邻域搜索。结论与单次划分的边长相比,新算法的边长估算更加合理、计算效率更高。Aim Based on the rate of point area different in space bounding boxes of different point clouds,to estimate the optimal edge length of grid.Methods According to the analysis of estimating methods for edge length of grid,a new algorithm of estimating optimal edge length based on point area is presented using second division.Results The algorithm′s scheme was given and the proposed method was applied to search the nearest point in the different point clouds.Conclusion Compared with the once division method,the edge length based on the proposed method is more reasonable and the speed of searching point is much faster.

关 键 词：空间包围盒 有点区域 二次分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]