本文为荷兰代尔夫特理工大学（作者：Mahshid Roumi）的硕士论文，共72页。

特征提取是各种图像处理应用中的关键功能。特征是一种图像性质，可以捕捉图像的某些视觉特性。纹理是许多图像类型的一个重要特征，它是图像中信息模式或结构的排列。纹理特征用于图像处理、遥感和基于内容的图像检索等不同的应用，这些特征可以通过多种方式提取。最常见的方法是使用灰度共生矩阵（GLCM）。GLCM包含图像相邻像素的二阶统计信息。纹理属性可以通过GLCM计算来了解关于图像内容的详细信息。然而，GLCM的计算量非常密集。本文设计并实现了一种用于GLCM快速计算的FPGA加速器，提出了一种基于FPGA的对称共生矩阵并行计算体系结构。实验结果表明，我们的方法可以将16个共生矩阵同时计算的处理时间提高2倍到4倍。

Feature extraction is a key function invarious image processing applications. A feature is an image characteristicthat can capture certain visual property of the image. Texture is an importantfeature of many image types, which is the pattern of information or arrangementof the structure found in a picture. Texture features are used in differentapplications such as image processing, remote sensing and content-based imageretrieval. These features can be extracted in several ways. The most common wayis using a Gray Level Co-occurrence Matrix (GLCM). GLCM contains the secondorder statistical  information ofneighboring pixels of an image. Textural properties can be calculated from GLCMto understand the details about the image content. However, the calculation ofGLCM is very computationally intensive. In this thesis, an FPGA accelerator forfast calculation of GLCM is designed and implemented. We propose an FPGA-basedarchitecture for parallel computation of symmetric co-occurrence matrices.Experimental results show that our approach improves 2x up to 4x the processingtime for simultaneous computation of sixteen co-occurrence matrices.

1 引言

2 数字图像与纹理特征

3 相关工作

4 FPGA实现

5 结论与未来研究展望

下载英文原文地址：

http://page2.dfpan.com/fs/0lc8j2b21029b16c8e4/ 

更多精彩文章请关注微信号：