面向城市地物分类的 SAR 图像纹理特征提取与分析

合成孔径雷达（ SAR）图像含有丰富的纹理信息，特别是进行城市地物分类时，纹理特征对于图像的解译具有重要的意义。本文对基于灰度共生矩阵和Gabor变换两种纹理特征提取方法进行了研究，将灰度和不同纹理特征组合应用于SAR图像城市地物分类，并以ALOS PALSAR影像为数据源进行了实验。通过对不同分类结果进行定性和定量分析，结果表明，引入纹理特征后的SAR图像分类结果要优于无纹理信息参与的分类结果，基于不同纹理特征组合的SAR图像分类结果要优于基于单一纹理特征的分类结果。

Extraction and Evaluation of Texture Features in SAR Images for Urban Land Cover Classification

Synthetic Aperture Radar （ SAR） images are rich in texture information , and texture features for image interpretation are of great significance , especially for urban terrain classification .In this paper , two kinds of texture feature extraction methods were used based on Gray Level Co -occurrence Matrix （ GLCM） and Gabor transformation , mainly to classify urban land cover of ALOS PAL-SAR images with the combination of gray scale and different texture features .The results of different classifications were analyzed qual-itatively and quantitatively , which showed that after the introduction of texture feature of SAR images classification result is better than that without the involvement texture classification , and different combinations of texture features based on SAR image classification re -sults are superior to the one based on a single classification result .