Hyperspectral remote sensing is a cutting-edge technology of acquiring land surface information, with a high resolution in spectrum. With the fast development of spaceborne and airborne sensors, hyperspectral data are more frequently and conveniently received. As a result, data storage, management and effective information extraction are becoming key challenges to hyperspectral remote sensing science and technology. Especially, automatic or semi-automatic information extraction is a predictabley trend for hyperspectral researches and applications.
In this dissertation, the intension and extention of the concept hyperspectral database and spectral data mining has been brought forward after a background research on hyperspectral remote sensing and database. Based on this conception, ground-object spectra database, remote sensing image database, data mining technology have been studied, while spatial data mining and image data mining technology are focused. A conceptual design with data mining and application oriented is provided basetd upon project experience and international study. The hyperspectral database has been built up and more than 6,000 sets of spectral data have been uploaded. There are spectra of rocks, minerals, waters, concretes, trees, wheats, soils and hyperspectral images in this database. With this hyperspectral database, some researches on spectral data mining and hyperspectral image data mining have been carried out. In view of hyperspectral remote sensing technology and application, there are some advangates of this dissertation as follows:
1. Based on hyperspectral database, new storage models of spectra and hyperspectral image has been put forward. A new design has combined data, methods and models in a whole database platform.
2. To integrage different sources of spectra in one database platform, a new conceptual data structure, dualistic-core star structure, has been brought forward.
3. Forward (from spectra to attributes) and backward researches have been impleted based on hyperspectral database. Automatic optimizations have been applied in band combination and band selection.
4. From a totally different view of hyperspectral image, a projection from hyperspectral images to relation tables has been built up to improve analysis and information extraction. The following key techniques are implemented in hyeprspectral database: Band selection based on Minium Length Model, Feature extraction based on Non-negative Matrix Fraction, Data compression and decompression based on Non-negative Matrix Fraction, Object Decection based on Support Vector Machine.
It is an important direction of international remote sensing technologies to develop aerial large array CCD digital remote sensing system. The development and digital progress of national aerial digital remote sensing technologies and photography devices will be greatly pushed about, and sharply descend the depends on foreign products of the same kind.
What’s more, the rapid improvement of China society and economics urgently requires large amount of high precision and high special resolution remote sensing images. Aerial large array CCD digital remote sensing system will just meet the requirement as it can obtain surficial information high availability.
In order to follow the currency, hyperspectral research group of Instituteof Remote Sensing Application, CAS pushes out a multi-mode airborne digital camera system, what is called MADC for short.
MADC system is combined with 3 digital cameras, which each has a large array CCD as 4K×4K. With different position and pose, the cameras can be combined to form different imaging modes, which are Wide Field, Multispectral, and Stereophotography mainly. MADC system can also be linked with Position and Orientation System (POS), and its cameras can be controlled to expose synchronistically associating with POS data according to regular time set by shutter control system. POS system can be seemed as the combination of GPS and IMU (Inertial Measurement Unit), and its data can show the spatial position and pose of aerial platform. To do geometric correction of remote sensing images aiding with POS data, it can get rid of the limitation of ground control points (GCP). So combined with normal aerial digital photographic devices, multi modes, multi spectral, and association with POS are the important characters of MADC.
This paper studies the imaging mechanism of each working mode of MADC, the calibration of MADC and the methods to correct MADC data with POS information for the fist time at home. The main contents include 7 aspects as showed below.
(1) Simplifying the geometric relationship of MADC’s cameras. By abstracting, the thesis gets the basic imaging mode of MADC and does detailed researches to the geometric of spacial projection. And the paper gets some important conclusions and inferences about MADC imaging, and even about normal mulit-angualer remote sensing.
(2) Analyzing the geometric factors, which are actually influencing imaging while MADC system working. The paper promotes interference models of atmosphere infraction, hypsography and earth curvature, and gets the conclusion that the interferences of atmospheric infraction and earth curvature are not neglected.
(3) Aiming at calibrating MADC system, the paper studies the distortion models of lens and does distortion correction of images. Basing on resection of single photo theory, the thesis gives methods to calibrate the pose of cameras’ optical axis and spacial position of pre-node of the lens by GCP and POS data.
(4) Researching how to relate POS data to image lines of linear CCD or images of array CCD, giving the relation methods for array CCD images, and resolving the problem of change mulit-center projection to single center projection.
(5) Studying the imaging effects of MADC multispectral mode and stereophotography mode, estimating the working results of optical filtering chips of MADC, and putting forward several possible stereophotography models.
Design and Implement application software platform is an important feature of thesystem 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 isthe 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 analysis, advanced tools, and integrated interface to connect with general spectral databases. The base architecture was specially designed and implemented to meet the requirements for the rapid preprocessing of imaging spectrometer data and the easy prototyping of algorithms. Some new methodologies for data analysis and processing were developed and applied to obtain results based on the architecture including mineral identification, agriculture investigation, and urban mapping. The key technologies of the Hyperspectral Remote Sensing Image Processing and Analysis System include the following: image tiling and processing, execution architecture, general data preprocessing, hyperspectral data calibration, hyperspectral data analysis, and image classification.
(5) The design of a new architecture for REMS. REMS provides the ability to extract the major characteristics of water resources, such as chlorophyll content, total suspended matter (TSM), and yellow substance. The most important components of REMS are presented in chapter Six, including tools for multi-resource data input/output, preprocessing, data visualization and mapping, environment information extraction, conventional image analysis, advanced tools for eco-environment modeling, and integrated interface to connect with general spatial and spectral database and water quality monitoring database. We established a REMS internet information publishing system REMS IS. Integrated with WebGIS and RS technology, REMS IS has four tiers and five subsystems, so that it can meet the specification of the B/S and C/S architecture. The four-tier architecture includes the following: the REMS application server, the middle data server, the web server, and the client. The REMS application server is composed with two important modules: the atmosphere correcting module and the water quality parameter inversion module 。。。。。。。。。...