[1]赵进超,朱颗东,申圳,等.基于Hadoop的图像文理特征提取[J].郑州大学学报(工学版),2015,36(04):109-113.[doi:10.3969/ j. issn.1671 -6833.2015.04.025]
　ZHAO Jin-chao,ZHU Hao-cong,SHEN Zhen,et al.Image Texture Feature Extraction Based on Hadoop[J].Journal of Zhengzhou University (Engineering Science),2015,36(04):109-113.[doi:10.3969/ j. issn.1671 -6833.2015.04.025]
点击复制

基于Hadoop的图像文理特征提取()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
36卷
期数:
2015年04期
页码:
109-113
栏目:
出版日期:
2015-08-31

文章信息/Info

Title:
Image Texture Feature Extraction Based on Hadoop
作者:
赵进超; 朱颗东; 申圳; 李红婢
郑州轻工业学院计算机与通信工程学院，河南郑州450002
Author(s):
ZHAO Jin-chao; ZHU Hao-cong; SHEN Zhen; LI Hong-chan
School of Computer and Communication Engineering, Zhengzhou University of Light Industry , Zhengzhou 450002, China
关键词:
Hadoop ;  ’Tamura纹理特征; 图像处理; 特征提取
Keywords:
Hadoop ; Tamura texture feature ; image processing ;  feature extraction
分类号:
TP301
DOI:
10.3969/ j. issn.1671 -6833.2015.04.025
文献标志码:
A
摘要:
随着数字图像规模的不断增加,图像纹理特征提取已成为制约数字图像处理性能的一个关键步骤.Hadoop是一个性能卓越的开源大数据处理云平台,其向用户提供了MapReduce ,HDFS等模块.首先对Hadoop平台、编程框架和Tamura纹理特征进行了介绍,然后将图像纹理特征提取过程在Hadoop平台上进行了实现.在这个过程中,每个Map任务对应一个图像文件,各节点可以同时提取集合内图像的纹理特征.实验表明:在图像数量较少和分辨率较低的情况, Hadoop 不同节点数量所用时间并无太大差异.在图像分辨率较高且数量较多的情况下 ,Hadoop平台表现出较高的计算效率.
Abstract:
With the increasing amount of digital image data,image texture feature extraction has become a keystep of digital image processing. As an excellent massive data processing and storage capacity of the opensource cloud platform,Hadoop provides a parallel computation model MapReduce,HDFS distributed file sys-tem module. In this paper,we firstly introduced Hadoop platform programming framework and Tamura texturefeatures. And then,the image texture feature extraction was carried out on the Hadoop platform. In theprocess,every Map task corresponds to an image file, every nodes work at same time.The comparison resultsshow that number of nodes have no influence about the processing time,when we have little images and theimage has low-resolution. On the contrary,Hadoop paltform is more effective.
更新日期/Last Update: