图像特征提取及识别过程

摘要

纹理特征是一种重要的视觉线索,是图像中普遍存在而又难以描述的特征。纹理分类与分割是图像处理领域一个经久不衰的热点研究领域,纹理特征提取作为纹理分类与分割的首要问题,一直是人们关注的焦点,各种纹理特征提取方法层出不穷。

本文在广泛文献调研的基础上,回顾了纹理特征提取方法的发展历程,分析了其研究现状,对纹理特征提取方法进行了较为全面的综述和分类，最后重点研究了基于灰度共生矩阵的图像纹理提取方法，研究如何有效地提取图像纹理特征来对图像进行描述，通过特征值来对图像进行识别。

灰度共生矩阵是一种简单有效的图像纹理特征描述方法，该方法的优势在于：它能利用了图像中像素相对位置的空间信息更加准确地描述图像的纹理，本文就是利用图像灰度共生矩阵的这一特性，从该矩阵中提取相应的统计参量作为纹理特征来实现对图像的识别。

关键字：灰度共生矩阵，纹理特征提取，图像识别

ABSTRACT

Texture is a kind of important visual clues in images , it is widespread but cannot easy to be described . Texture classification and segmentation is a enduring popular research field in image processing area. Texture feature extraction has been the focus of attention,due to its priority to texture classification and image segmentation. all sorts of texture feature extraction methods has been emerged in endlessly.

On the basis of extensive literature investigation, we review the texture feature extraction methods, analyze the development of the research status of the texture feature extraction methods and make a comprehensive review of its classification . Finally ,based on gray symbiotic matrix image problem extraction methods,we research how to effectively extract image texture feature described by the image characteristic value to image recognition.

Graylevel co-occurrence matrix is a simple and effective image texture description method.This method's advantage is: it can use the image pixels relative positions of the spatial information more to accurately describe the texture image.This paper use the graylevel co-occurrence matrix of the properties to extract statistics from the matrix corresponding as texture feature parameters to realize image recognition.

KEY WORDS: graylevel co-occurrence matrix, texture feature extraction, image recognition

目录

前言............................................... 错误!未定义书签。第1章图像纹理特征概述............................. 错误!未定义书签。

1.1 发展与现状.................................. 错误!未定义书签。

1.2纹理的有关定义............................... 错误!未定义书签。第2章纹理特征提取方法............................. 错误!未定义书签。

2.1 纹理特征提取分类............................ 错误!未定义书签。

2.2 统计家族的灰度共生矩阵...................... 错误!未定义书签。第3章图像特征提取及识别过程....................... 错误!未定义书签。

3.1 系统流程图................................. 错误!未定义书签。

3.2 灰度共生矩阵定义............................ 错误!未定义书签。

3.3 四个方向灰度共生矩阵的生成.................. 错误!未定义书签。

3.4 纹理特征参数的介绍.......................... 错误!未定义书签。第4章算法实现..................................... 错误!未定义书签。第5章对此次设计的总结与展望....................... 错误!未定义书签。

5.1 设计中遇到的问题............................ 错误!未定义书签。

5.1.1 对纹理的理解问题....................... 错误!未定义书签。

5.1.2 程序调试方面的问题..................... 错误!未定义书签。

5.1.3 论文攥写的问题......................... 错误!未定义书签。

5.2 总结与展望................................. 错误!未定义书签。致谢词：............................................ 错误!未定义书签。参考文献：.......................................... 错误!未定义书签。附录：.............................................. 错误!未定义书签。