混合像元分解方法的研究

混合像元是遥感领域研究的热点，也是难点。混合像元问题不仅影响地物识别和分类精度，而且是遥感技术定量化发展的重要障碍。为了提高遥感应用的精度，就必须解决混合像元的分解问题。该文首先介绍了混合像元的产生机理，然后阐述了时下几种流行的混合像元分解模型：线性模型、概率模型、几何光学模型、随机几何模型和模糊分析模型，并对其中几种常用模型的优缺点及其适用性进行总结和讨论．对不同模型之间的相似和差异性进行比较分析。

Research on the Methods of Unmixing the Mixed Pixels

Mixed pixels is hot and difficult topic in remote sensing research area. The topic of mixed pixeis not only has effect on distinguishing ground objects,but also is an barrier to the quantitative development of remote sensing technology.In order to improve the precision of remote sensing application, resolving the problem of unmixing the mixed pixels should be done.This paper first exposits the mechanism of mixed pixels,second, elaborate some prevalent models of unmixing the mixed pixels,such as Linear Model,Probabilistic Model,Geometric-Optical Model,Stochastic Geometric Model and Fuzzy Model,then, then summarize and discuss several commonly used models＇ advantages and disadvantages, last,comparatively analysis the differences and similarities between different models.