【作者】 王志平；

【导师】 杨建峰；

【作者基本信息】 中国科学院研究生院（西安光学精密机械研究所） ， 光学工程， 2007， 硕士

【摘要】 针对西安光机所研制的星载高光谱成像仪和月球干涉成像光谱仪光谱应用技术展开研究。首先讨论了数据预处理技术，其次针对星载高光谱成像仪进行了地物覆盖光谱分类技术和目标探测技术研究。最后针对月球干涉成像光谱仪进行了矿物质混合光谱分解技术研究。讨论了数据预处理技术，主要是辐射校正和反射率反演方法。利用此方法处理星载高光谱成像仪飞行实验数据，得到了可信的反射率光谱图像。利用新的基于端元的地物识别分类技术处理了飞行实验数据，得到了较好的分类结果。利用BandMax和SAM相结合的技术处理飞行实验数据进行目标探测，假定目标(汽车)大部分被探测到。针对月球干涉成像光谱仪进行混合光谱分解技术研究，讨论了高光谱解混合的三个步骤，重点介绍解混合(inversion逆问题求解)这一步骤。最后对AVIRIS获取的Cuprite地区的矿物质进行混合光谱分解，结果可靠。为即将获取的月球高光谱影像分析提供了一种有效途径。更多还原

【Abstract】 Hyperspectral data application for land cover and mineral analysis is discussed in this paper. Four main parts are included: Hyperspectral data preprocessing, Land cover classification using an endmember-based method, Target detection and mineral mapping using spectral unmixing.Hyperspectral data preprocessing is discussed, including radiation correction and atmospheric correction. The flying data is processed using this preprocessing method, and they get a credible reflectance image.A new method for land cover mapping, endmember-based identification and classification, is used to process the flying data. They get a satisfying classification result.The interesting targets (cars) are detected using BandMax and SAM. They get a satisfying detection result.Mineral mapping for the lunar interference imaging spectrometer is studied. Three steps for spectral unmixing are discussed. They process the hyperspectral image of the Cuprite derived by AVIRIS using the spectral unmixing method for mineral mapping. This method can be used in lunar hyperspectral image processing future.更多还原