基于海量数据的二维凸包快速生成算法  

Fast Algorithm for Generating Two-dimensional Convex Hull Based on Mass Data

作　　者：马骏[1] 蔺东杰 凌广明[2] 

机构地区：[1]河南大学计算机与信息工程学院,河南开封475004 [2]河南大学软件学院,河南开封475004

出　　处：《计算机技术与发展》2017年第2期42-45,共4页Computer Technology and Development

基　　金：国家自然科学基金资助项目(61202098)

摘　　要：凸包算法是计算机几何的基本问题之一,在很多领域应用广泛。传统的凸包生成算法在处理大容量数据时,表现出的时间复杂度相对较高而且凸包生成速率较低,已经不能满足实际海量数据的需求。为解决这一问题,提出了一种面对海量数据的快速凸包生成算法。该算法通过对散乱点集分区、一遍扫描排序,确定散乱点集边界,快速处理边界点集中处于共线的点等一系列预处理操作,快速排除凸包内部的点,缩小了问题规模,避免了对不在凸包上的点集的扫描处理,明显地缩短了凸包的求取时间,可保证最小凸包的快速生成。该算法极其简单,时间复杂度较低,理论上可达到o(nlogn),有利于凸包生成速度的提高。与传统算法进行了同步对比实验,结果表明,该算法运行有效性较好,且具有较好的应用前景。The convex hull algorithm is one of the fundamental problems in computer geometry and has been widely used in many fields.Traditional convex hull generation algorithm in dealing with large amounts of data shows high time complexity relatively and the lower rate of convex hull generation,which has been unable to meet the needs of actual data.In order to solve this problem,a fast convex hull algorithm of massive data in the face is proposed. Through a series of pre- processing operations of determining the scattered point set boundary and fast processing of boundary points concentrated in collinear points by partition and over scan sort of scattered point set,the algorithm quickly rules out the points inside the convex hull and reduces the size of the problem,to avoid the processing of assemblies that are not in the convex hull of point scanning,significantly shortening the calculating time of the convex hull,which can ensure the rapid generation of minimum convex hull.The algorithm is extremely simple,with lowtime complexity,achieving o（ nlogn） theoretically. It is conducive to improve the speed of convex hull.Compared with the traditional algorithm,the experimental results showthat the proposed algorithm is effective and has good application prospects.

关 键 词：凸包 海量 平面点集 预处理 排序 快速 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]