中文摘要：

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.

英文摘要：

With the increasing amount of digital image data,image texture feature extraction has become a key step of digital image processing. As an excellent massive data processing and storage capacity of the open source cloud platform,Hadoop provides a parallel computation model MapReduce,HDFS distributed file system module. In this paper,we firstly introduced Hadoop platform programming framework and Tamura texture features. And then,the image texture feature extraction was carried out on the Hadoop platform. In the process,every Map task corresponds to an image file,every nodes work at same time. The comparison results show that number of nodes have no influence about the processing time,when we have little images and the image has low-resolution. On the contrary,Hadoop paltform is more effective.