基于纹理特征的 SAR 影像各拉丹冬冰川信息提取

摘要

Glacier area is an important parameter to monitor glacier change information.Based on the specific texture features of glac-ier region, the paper selected a time interval of 35 days of ENVISAT ASAR images, in use of GLCM to generate texture features, then used supervised classification method by band combination, and then extracted area of glaciers;while on the basis of Landsat TM opti-cal image to evaluate the use of texture feature extraction accuracy of the results.Research shows that:the use of SAR images based on texture feature extraction methods is feasible to extract information of glacier.%冰川面积是监测冰川变化信息的重要参数。本文以各拉丹东地区为例，根据冰川区域特有的纹理特征，选取时间间隔为35天的ENVISAT ASAR干涉对，利用灰度共生矩阵提取纹理特征，通过波段组合进行监督分类，进而提取研究区冰川面积。同时以Landsat TM光学影像为依据，评价利用纹理特征提取结果的精度。研究表明：基于纹理特征并利用SAR影像提取冰川面积的方法是可行的，为提取冰川信息提供了又一可靠手段。