基于线性光谱模型的混合像元分解方法与比较_陈峰精选.pdfVIP

遥感信息! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 2010.4 基于线性光谱模型的混合像元分解方法与比较 , 陈峰 , 邱全毅 , 熊永柱 , 黄少鹏 ( , , 361021; , 514015; , , 48109- 1005 ) :线性光谱模型是目前解决城市中等空间分辨率遥感( 如Landsat) 中存在的混合像元问 的简单有效 的策略本实验以广州区域为研究区, 利用ENVI/ IDL 影像处理和开发平台对4 种混合像元线性光谱分解方法 进行了对比, 即无约束条件法带部分约束条件法普通带全约束条件法和带全约束条件的可变端元法结果表 明, 普通带全约束条件法和带全约束条件的可变端元法的分解结果比无约束条件法和带部分约束条件法的分解 结果合理, 均方根误差明显要小; 同时, 带全约束条件的可变端元法要优于普通带全约束条件法光谱归一化处 理则对不同分解方法带来不同的影响,应依据实际需要采取合适的光谱处理方式 : Landsat;混合像元; 线性光谱模型 doi : 10. 3969/ j. issn. 1000- 3177. 2010. 04. 005 : TP79 ! ! : A ! ! : 1000- 3177( 2010) 110- 0022- 07 Pixel Unmixing Based on Linear Spectral Mixture Model : Methods and Comparison , ! ! CH EN Feng , QIU Quanyi , XIONG Yongz u , H UANG S aopeng ( Key Lab of Urban Environment and Health,Instituteof Urban Environment, Chinese Academy of Sciences,Xiamen 361021; ep artment of Geograp hy, Jiaying University ,Meizhou 514015; ep artment of Geological Sciences, University of Michig an, Ann Arbor,MI 481091005, USA ) Abstract: At present, Linear Spectral Mixture M odel ( LSMM) as been considered as a simple but effective way to extract useful information from t e mixed pixel w ic is t e problem confronted in t e application of spatial mediumresolution remote sensing images ( e. g . Landsat) to study t e urban environment. Taking Guangz ou as a study region,four unmixture approac es based on t e concept of LSMM w ere developed and compared using ENVI/ IDL,w ic is a popular platform for image process ing and code development. Unmixture approac es in t is study are termed unconst rained, partialconstrained, g