高光谱图像端元提取算法研究进展与比较  

The development and comparison of endmember extraction algorithms using hyperspectral imagery

作　　者：李二森[1,2] 朱述龙[1] 周晓明[1] 余文杰[1] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052 [2]矿山空间信息技术国家测绘局重点实验室,河南焦作454000

出　　处：《遥感学报》2011年第4期659-679,共21页NATIONAL REMOTE SENSING BULLETIN

基　　金：矿山空间信息技术国家测绘局重点实验室(编号:KLM200904)~~

摘　　要：高光谱图像中混合像元的存在不仅影响了基于遥感影像的地物识别和分类精度,而且已经成为遥感科学向定量化方向发展的主要障碍。本文分析和研究了现有的典型端元提取算法,在此基础上,对这些算法进行归纳总结,从是否假定纯像元存在角度将其分为两类:端元识别算法和端元生成算法,并就两种分类方法选取了具有代表性的6种典型端元提取算法:N-FINDR、VCA、SGA、OSP、ICE和MVC-NMF算法进行分析和实验。通过对这6种方法的实验比较,得出两种端元提取分类方法的优点与不足,并对今后的研究工作提出展望。The mixels in the hypersepectral images not only influence the accuracy of target detection and classification, but also greatly hinder the development of quantitative remote sensing. The typical endmember extraction algorithms now available are analyzed and summarized. These algorithms can be classified into two types based on the hypothesis of the existence of the pure pixels： endmember identiflcation algorithm and endmember generation algorithm. Six endmember extraction algorithms, including N-FINDR, VCA, SGA, OSP, ICE and MVC-NMF, are introduced and compared using experimental data, which further show their advantages and disadvantages. With results of various methods, the future perspective is proposed for further study.

关 键 词：高光谱图像 混合像元 线性光谱混合模型 端元 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]