基于函数近似原理的遥感数据分类算法

将人不可及的观测通称为遥感。高光谱卫星遥感数据是由人造卫星获得的一系列地表物体电磁波信号反射值。高光谱数据信息含量大,其各电磁波波段波长间隔最小可在10nm左右6]。文章所介绍的聚类算法是一种专门用于处理高光谱卫星遥感数据的数据挖掘新算法。该算法以函数近似原理为理论基础,考虑了温度变化对于遥感数据的干扰影响。将该算法与ENVI软件提供的督导类分类算法进行比较,实验结果表明在样本数据较少情况下该算法的性能和可靠性都具有相当的优势。为了表述方便将文章提到的聚类算法简称为＂函数聚类＂算法。

A Classified Algorithm Based on Function Approximate Theory for Remote Sensing Data

Remote sensing is traditionally defined as a technology for the acquisition of data about a target without touching it.Remote sensing data is a kind of information that illustrated the reflectance value of electromagnetic wave effect on the earth surface.The spectral resolution of hyperspectral data is higher than multispectral data.In hyperspectral data,the spectral resolution is usually small as 10nm.Hyperspectral data have a large information6].In this paper,a new algorithm has been proposed which is sp...