基于高光谱数据(Hyperion)混合像元分解的研究

本文以高光谱(Hyperion)数据为数据源,借助遥感手段对数据进行坏线修复、大气校正和几何校正等处理,通过有约束条件和无约束条件线性混合模型对像元对研究区域进行分类研究.研究结果表明:无约束条件和有约束条件进行混合像元的分类精度分别为90.1%和90.4%,两者的分类精度相差不是太大;就裸地而言,有约束条件的线性像元混合分解的精度要远高于无约束条件;从冗余的数据中提取有用信息和如何利用高光谱遥感的今后遥感图像分类研究的难点和热点.

Study On Decomposing Mix Element Based On Hyperion Data

This paper uses hypectrum in order for atmospheric correction,geometric correction,and so on,and uses re- straint linearity mixed model and nonrestraint linearity model to classify the pixels and the area we analyze.The research indicates that,the precision of restraint classification is 90.1%,and that of nonrestraint classification is 90.4%,the precision of restraint and nonrestraint,analyst is not too much;To the bare land,the precision of restraint analyst is much higher than that of nonrestraint analyst.In future.,to get useful information from redundant data and how to use hyperspectral remote sensing will be the difficulty and hotspot in the image classification analyst of remote sensing.