中国科学院遥感与数字地球研究所-高光谱遥感应用技术研究室推荐内容
【2012硕】基于FISS 的岩矿多维高光谱数据库研究
|
【作者】 覃环虎 【导师】 童庆禧；张立福 【 学位年度 】 2012 【论文级别】 硕士 【关键词】 高光谱数据库，数据存储模型，多维数据库，FISS 【Key words】 Hyperspectral database, Data storage model, Multidimensional database, FISS 【中文摘要】 随着各种高光谱传感器的成功研制，高光谱遥感技术正得到日益广泛的应用，其所积累的数据量也与日俱增。如何有效的管理和共享海量高光谱数据，并从海量数据中获取有益的信息，是遥感科学工作者面临的一个重要挑战。建设高光谱数据库是解决数据有效管理和共享的方法，而多维数据库技术和多维分析技术的发展，将为从海量数据中获取有益信息提供技术基础。 本文以自主研发的地面成像光谱辐射测量系统（Field Imaging Spectrometer System,FISS）所采集到的高光谱数据为研究对象，以建立岩矿多维高光谱数据库为最终目标，开展了两个方面的研究工作：一是分析、设计与开发了FISS 高光谱数据库原型系统，二是基于多维数据库的理论，以FISS 高光谱数据库存储的数据为数据源，探讨了岩矿多维高光谱数据库的基本概念、数据库设计以及多维分析工具集设计。 论文的主要成果和结论如下： 1. 提出了以GeoRaster 对象为核心的FISS 数据存储模型，在该存储模型的基础上，对基于存储过程的数据入库方法和基于动态数据表转换的光谱数据获取方法进行了深入研究。研究结果表明，该存储模型和数据入库方法能够有效的支持高光谱数据图谱合一的存储。 2. 在Arc Engine 开发工具和C#的支持下，研究了FISS 合成影像的显示方法和光谱曲线的显示方法。实验结果表明，数据获取方法、影像显示方法和光谱曲线显示方法的完美结合，有效的实现了高光谱数据图谱合一的显示。 3. 在成功研究数据存取技术、影像和光谱显示方法的基础上，开发实现了FISS 高光谱数据库原型系统，在系统中集成了一些具有高光谱特色的光谱分析方法。该系统基本具备了对高光谱数据进行管理、显示和分析的功能。 4. 以多维数据库理论为基础，研究了FISS 岩矿多维高光谱数据库的基本概念，初步设计了它的系统组成结构，提出了嵌套的岩矿多维数据结构，构建了岩矿多维数据的雪花模型，研究了其多维分析工具集的组成，认真设计了每一个多维分析操作的具体功能，并展示了多维分析的效果。这些基本研究为下一步工作的开展奠定了坚实的基础。
【Abstract】 Although hyperspectral remote sensing technology is advancing our understanding in various fields, it also brings us data redundancy. It is really a great challenge to manage and share hyperspectral remote sensing data effectively and obtain useful information from such massive data. We believe the establishment of a database for effective management and sharing of hyperspectral remote sensing data is really necessary. Meanwhile, the development of multidimensional database technology and multidimensional analysis technology will lay the foundation for obtaining useful information from massive data. This study aims at the establishment of a multidimensional hyperspectral database for rock and mineral based on the hyperspectral data collected by the Field Imaging Spectrometer System (FISS).This study includes two parts:1) design and develop the prototype system of FISS Hyperspectral Database (FHD); 2) probe into the basic concepts and design of the multidimensional hyperspectral database for rock and mineral (MDHDR&M) as well as the design of multidimensional analysis tools based on multidimensional database theory. Main results and conclusions are summarized as follows: 1. FISS data storage model which key is GeoRaster is proposed. Extensive researches were carried out on data import method based on procedure and data extract method transformed through dynamic data table. The experimental results indicate that this model and data import method work well at the integrated storage of both image and spectrum of hyperspectral data. 2. Researches about the visualization of the FISS image and the visualization of FISS spectral curve are based on Arc Engine and C#. The result shows that flawless combination of data obtained method, image visualization method and spectral visualization method can support the integrated visualization of spectrum and image of hyperspectral data effectively. 3. A prototype system of FHD with several spectral analysis methods is developed. This prototype system is composed of such functions as management, visualization and analysis. 4. This study redefines the basic concept of MDHDR&M and preliminarily designs its system structure. This model also brings in a multi-dimensional nested data structure to establish a snowflake schema. Besides, the compositions and detailed functions of multidimensional analysis tools are analyzed. Some results of multidimensional analysis are showed. All these lay the foundation for the further implementation of hyperspectral multi-dimensional database for rock and mineral.
(责任编辑：admin) |
