高光谱遥感数据挖掘若干基本问题的研究

面向高光谱遥感信息的特点，分析了高光谱遥感数据挖掘的形成和作用，在构建其框架体系与处理流程的基础上。探讨了可以发现的知识类型和典型的挖掘模式，并分析了一些主要挖掘算法和关键技术。最后对高光谱遥感数据挖掘潜在的应用方向进行了探讨。

Some Key Issues of Hyperspectral Remote Sensing Data Mining

The formation and role of Hyperspectral Remote Sensing Data Mining (HRSDM) are analyzed oriented to the characteristics and demands of hyperspectral information processing. The framework and processing flow of HRSDM are proposed at first. The knowledge that can be discovered from hyperspectral RS information includes spectral signatures, spatial rules, knowledge about genesis and diagnosis, spectral knowledge among different bands, dynamic evolution knowledge and isolated point identification. Five useful DM modes are standard spectral database mining, spectrum-based mining, spatial dimension-based mining, spatial and spectral dimension mining and mining operations with support of auxiliary and background data. By analysis and experiments, some algorithms proved effective to HRSDM include association rule mining, clustering and artificial neural network (ANN), rough set and fuzzy theory, decision tree and data cube. Finally, some potential applications of HRSDM including typical information extraction, quantitative RS, RS inversion, image classification, mixed pixel decomposition, feature extraction and selection and optimal band combination are put forward.