基于形态学的土地荒漠化遥感图象纹理分析

地貌环境不仅影响土地荒漠化的速率，而且制约土地荒漠化的方式和空间分布规律。地貌类型、特别是地形起伏程度对荒漠化的影响首先表现在荒漠化方式上，其次表现在荒漠化强度上。遥感影象本身通常都包含着灰度和空间信息。本文利用数学形态学进行纹理分析，从腐蚀运算结果抽取纹理特征、利用形态学运算抽取分形参数作为纹理特征等，从而提取地物的空间结构信息。利用纹理结构参与土地荒漠化遥感分类，大大提高了土地荒漠化分类的精度。

Morphology-based Texture Analysis for Remote Sensing Image of Desertification

The landforms environment affects not only the rate of desertification, but also the mode and distribution patterns. The influence to desertification of the landforms type especially the topographic fluctuation was presented firstly to the desertification mode and secondly to the desertification strength. Remote sensing images contain gray and space information. In this paper, mathematical morphology was used to process texture analysis. From the corrosion and morphology calculation results texture feathers were extracted to describe the space structure information of the ground objects. The experiment shows that the texture structure information added in desertification classification can raise the classification accuracy greatly.