基于小波变换和模糊积分的遥感图像分类

遥感影像分类是遥感影像数字处理的一个重要内容。针对实地数据缺乏背景下的遥感影像分类研究,文中提出了一种基于小波变换和模糊积分的遥感影像分类方法。该方法针对遥感影像的纹理结构与成像条件的相关性,分别对多光谱影像的三个通道纹理图像做聚类,然后采用模糊积分技术融合三者结果得到最终的分类图像。实验结果表明本文方法的分类结果要优于任何单通道的聚类结果和光谱聚类结果,且其分类连通性优于有监督的光谱分类结果。

Remote Sensing Image Classification Based on Wavelet and Fuzzy Integral

Remote sensing image classification is an important content in remote sensing image processing research.Under the circumstance of lack exact field data,a remote sensing image classification method based on wavelet and fuzzy integral is presented in this paper.From the relation between texture and imaging,texture image of each band is classified with their results fused by Choquet fuzzy integral.Experimental results show that the proposed method outperforms any one using single band as well as that of spectrum feature,it also provides a better connectivity than the classification of supervised Kmeans using spectrum feature.