Research On The Key Technologies Of Hyperspectral Remote Sensing Environment Monitoring System And Its Applications To Water Resources

Design and Implement application software platform is an important feature of the system project of the Earth Observing Plan. In this thesis, we focus our research on Hyperspectral Remote Sensing Environment Monitoring System and its applications to water resources.To meet the requirements for water quality monitoring in China, a Remote-sensing Environmental Monitoring System (REMS) is introduced. REMS is the first integrated system in developed for multi-resource, multi-temporal, and multi-thematic data processing and data analysis, and for distributing products for the monitoring of inland water pollution using remote sensing technology. REMS provides the ability to quickly extract the major characteristics of water resources, such as chlorophyll content, total suspended matter (TSM), yellow substance, and blue algae distribution. In order to improve the precision of water parameter extraction, new algorithms and functions are also developed and integrated into the REMS software platform. We select the Taihu Lake in Jiangsu Province, China, as the research area. Finally, we discuss some field applications constructed based on REMS.My dissertation includes the following:(1) The introduction of hyper-spectral data algebra structure to hyperspectral algebra structure analysis (HASA).(2) The establishment of semi-empirical and analytical models to extract the major characteristics of water resources such as chlorophyll content, total suspended matter (TSM), Yellow Substance, and CDOM. The analytical models are based on bio-optical models and radiation-transfer theory, where the optical properties and water quality parameters have distinct physical meaning and universal applicability.(3) The introduction of fractal-based image compression algorithms using wavelet transformation for hyperspectral images in Chapter four. This algorithm is superior to other traditional compression methods because it has high compression ratios, good image fidelity, and requires less computation time. I present a fast fractal image coding methodology based on wavelet decomposition. The subimage blocks are improved by Pyramidal Haar transform. Furthermore, the improvement of the Quadtree partition is also discussed. My simulations show that the coding time is 100 times faster when the same coding result is retained. The HV and Quadtree partitioning and the domain-range matching algorithms have also been improved to accelerate the encode/decode efficiency.(4) The design of the most important components of the Hyperspectral Remote Sensing Image Processing and Analysis System, including tools for input/output, preprocessing, data visualization, information extraction, conventional image...