matlab纹理分析,森林遥感图片的纹理分析(MATLAB)☆

摘要

遥感技术不断提高，森林遥感图像所含信息越来越多，仅用光谱信息无法将其区分开，而用纹理特征分析对于在图像的识别起着非常重要的作用，因此遥感影像的纹理分析已经成为一种重要的提高遥感影像分类精度的手段。本文以森林遥感图片为研究对象，学习纹理分析的不同方法，选择合适且简单的方法对森林遥感图像进行纹理分析。首先，针对森林遥感图像的特点并结合现行纹理分析的不同方法，选择适于描述森林纹理的灰度共生矩阵方法，并在此基础上提取纹理特征参数。其次，对图像进行平滑，选择一个均值模版让图像各点间的距离更加接近使图像平滑。最后，利用k-means算法让平滑后的图像结合提取的特征值对图像各点进行聚类分析，把相近的的点归在一起。本论文的算法用MATLAB来实现，实验结果表明该论文所用的分析方法对森林遥感图像的纹理分析具有一定的可靠性。

关键词：森林遥感图片；纹理分析；灰度共生矩阵；纹理特征值；k-means算法；聚类分析

Abstract

Constantly has remote sensing technology improved, forest remote sensing image which contains more information. It is impossible to separate it only with spectral   information, whereas texture analysis plays a important role in identification of image. Consequently texture analysis of remote sensing image has become a critical approach on improving classification accuracy of remote sensing image. The paper, consider forest remote sensing image as a study object and learn the different methods of texture analysis, choosing a suitable and simple method to analyze texture of forest sensing image. First of all, select the gray level co-occurrence matrix approach which is suitable to describe texture of