Hyperspectral And High Spatial Resolution Remote Sensing Information Fusion Technology Research

In this thesis the data processing and analysis of high resolution satellite and hyperspectral imagery are researched. The aim of this dissertation is put on the image fusion in the high resolution imagery such as IKONOS, QuickBird and to develop methods and models for hyperspectral data processing and analysis.The principles and technical methods of some popular image fusion models that include Color transform , PCA transform , HPF transform , wavelet transform are first described and analyzed, then on the basis of wavelet transformation algorithm and feature of high resolution imagery, an area feature wavelet package transformation algorithm has been put forward. Furthermore, evaluation for fusion algorithms is introduced. A decision-level fusion method based on rough set theory for multisource and huge information has been given, which can make selection of fusion methods easy and efficient.The principles of hyperspectral imaging systems and the methods for hyperspectral data processing were studied. First base upon the sensor and optical models ,the various methods for reflectance conversion were discussed . By using the empirical line method,reflectance conversion of shanghai hyperspectral data were made. Then some key technologies which including hyperspectral data calibraton and parameterization, hyperspectral image classification and fusion ,bands selection , spectral mathing are analyzed, implemented and validated by experiments. In case of so much spectral bands and huge data,a data-manage model of spectral library base on oracle database is developed. Lastly a prototype system MRSIFS has been designed and realized.