基于小波变换特征的遥感地貌影像纹理分析和分类  

REMOTE SENSING RELIEF IMAGE TEXTURE ANALYSIS AND CLASSIFICATION BASED ON WAVELET TRANSFORM FEATURES

作　　者：朱长青[1,2] 杨启和[1,2] 朱文忠[1,2] 

机构地区：[1]郑州测绘学院 [2]北京大学视觉与听觉国家重点实验室

出　　处：《测绘学报》1996年第4期252-256,共5页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金

摘　　要：本文基于图像的正交小波变换特征，研究了遥感地貌纹理影像的特征提取和分类方法，并对２５幅地貌影像进行了分类。结果表明，所述方法不仅对同一分辨率的影像有较高的分类正确率，而且对不同分辨率的影像也有较高的分类正确率。Based on orthogonal wavelet transform features of image, in this paper, the method of feature selection and classification was studied for remote sensing texture images. The classification experiments had been carried out for 25 relief images. The experiment results showed that the high accuracy could been obtained by the method not only for the images with the same resolution or different resolutions, but also for the images that test samples and training samples come from different images.

关 键 词：地貌 影像纹理 遥感 小波变换 分类 

分 类 号：P931[天文地球—自然地理学]