计算机应用 ›› 2015, Vol. 35 ›› Issue (5): 1296-1301.DOI: 10.11772/j.issn.1001-9081.2015.05.1296

• 先进计算 • 上一篇    下一篇

多核CPU下的K-means遥感影像分类并行方法

吴洁璇, 陈振杰, 张云倩, 骈宇哲, 周琛   

  1. 江苏省地理信息技术重点实验室(南京大学), 南京 210023
  • 收稿日期:2014-12-15 修回日期:2015-01-11 出版日期:2015-05-10 发布日期:2015-05-14
  • 通讯作者: 陈振杰
  • 作者简介:吴洁璇(1991-),女,浙江台州人,硕士研究生,主要研究方向:高性能地理计算、数据挖掘; 陈振杰(1974-),男,陕西宝鸡人,副教授,博士,主要研究方向:GIS算法、高性能地理计算; 张云倩(1992-),女,江苏扬州人,硕士研究生,主要研究方向:空间数据挖掘;骈宇哲(1989-),男,陕西铜川人,硕士研究生,主要研究方向:高性能地理计算; 周琛(1990-),男,江苏宿迁人,博士研究生,主要研究方向:高性能地理计算.
  • 基金资助:

    国家863计划项目(2011AA120301);国家科技支撑计划项目(2012BAH28B02).

Utilizing multi-core CPU to accelerate remote sensing image classification based on K-means algorithm

WU Jiexuan, CHEN Zhenjie, ZHANG Yunqian, PIAN Yuzhe, ZHOU Chen   

  1. Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology (Nanjing University), Nanjing Jiangsu 210023, China
  • Received:2014-12-15 Revised:2015-01-11 Online:2015-05-10 Published:2015-05-14

摘要:

针对海量遥感影像快速分类的应用需求,提出一种基于K-means算法的遥感影像并行分类方法.该方法结合CPU下进程级与线程级模式的并行特征,设计融合进程级与线程级并行的两阶段数据粒度划分方法和任务调度方法,在保证精度的基础上实现并行加速.利用大数据量的多尺度遥感影像进行实验,结果表明:所提并行方法可大大减少遥感影像的分类时间,取得了良好的加速比(13.83),并可达到负载均衡,从而解决了大区域遥感影像快速分类的问题.

关键词: K-means算法, 并行计算, 负载均衡, 数据粒度划分, 消息传递接口, OpenMP

Abstract:

Concerning the application requirements for the fast classification of large-scale remote sensing images, a parallel classification method based on K-means algorithm was proposed. Combined the CPU process-level and thread-level parallelism features, reasonable strategies of data granularity decomposition and task scheduling between processes and threads were implemented. This algorithm can achieve satisfactory parallel acceleration while ensuring classification accuracy. The experimental results using large-volume and multi-scale remote sensing images show that: the proposed parallel algorithm can significantly reduce the classification time, get good speedup with the maximum value of 13.83, and obtain good load-balancing. Thus it can solve the remote sensing image classification problems of the large area.

Key words: K-means algorithm, parallel computing, load balancing, data granularity decomposition, Message Passing Interface (MPI), Open Multi-Processing (OpenMP)

中图分类号: