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
【2003硕】网络环境下高光谱数据库构建及其应用实践
|
【作者】 张雄飞 【导师】 童庆禧；郑兰芬；张兵 【 学位年度 】 2003 【论文级别】 硕士 【关键词】 高光谱数，据库 网络，ORACLE 【Key words】 Hyperspectral Database, Internet , ORACLE 【中文摘要】 高光谱遥感作为遥感的前沿技术，正在得到日益广泛的应用，其数据积累日益丰富。随着定量化遥感渐渐成为关注的核心。很多领域都越来越迫切的要求建立一个存放高光谱数据，功能性强，使用方便，易扩展的，与本领域应用相结合的数据库系统。与此同时，高光谱数据却有着其与众不同的特点，一方面由于采用仪器的不同，造成波段数各不相同；一方面其一个对象对应各种属性数据、图像数据以及许多波段的光谱数据，数据量大，关系复杂；同时高光谱数据还具有图谱合一的显著特点，所有这些决定了原有的光谱数据库系统已经不能满足高光谱数据研究、应用的需要。 本文总结了前人的研究成果，研究了目前光谱数据库的发展现状，根据高光谱数据应用的需要，在国际和国内首次提出了高光谱数据库的概念，它无论在存储地面成像光谱辐射计数据方面还是在处理图像光谱数据方面都具有十分重要的现实意义。高光谱数据库不但继承了原有光谱数据库系统存储、显示光谱及其相关属性数据的特点，而且为图像光谱数据设计了存储结构，将之妥善的存储、显示，从数据库的角度反映出了高光谱数据图谱合一的特点，使数据结构更为紧凑，将图像与光谱融为一体，为将来海量高光谱数据的挖掘做好了准备。 一个高光谱数据库系统将主要由以下几部分组成：高光谱图像样本库系统；光谱数据库辅助系统；高光谱数据分析系统；带有数据挖掘功能的数据仓库以及前台界面系统。本文在提出这个组织结构的同时，实现了一个高光谱数据库原型系统。主要包括：基于网络的光谱数据库系统，这是原有光谱数据库系统的发展，在应用的数据平台上发展到了大型的ORALCE数据平台，随着全系统的建成，它还具有了数据分析系统以及网络化的前台发布。同时还将作为高光谱数据库系统的辅助子系统；基于网络的数据分析系统，这是为了扩展数据库应用而建立的具有基本高光谱数据分析功能的子系统，它将以数据库中存储的各种数据为对象，进行数据处理，丰富数据库的功能；基于网络的光谱图象子系统，这是整个高光谱数据库系统的灵魂所在，它从数据库层面上实现了高光谱数据图谱合一的特性，将光谱与图像融为一体，从图像中可以直接提取光谱。这种将空间信息、光谱信息集成为一体的数据库系统，必将发挥其重要的作用；前台发布系统，在信息时代，网络化已经成为发展的趋势，在网络上实现整套系统的发布将有利于促进数据的交流与共享，开拓高光谱应用的范围和前景。本文以高光谱数据库系统概念为核心，围绕这几个子系统的建立，详细讲述了技术路线、基本框架，各子系统设计时所遇到的特殊问题，以及解决他们采取的方法和步骤，乃至最终实现的结果，全面介绍了一个相对比较完整的基于网络的高光谱数据库原型系统的建设过程。 同时，在设计、解决问题和实现过程中，从理论上为高光谱数据库系统总结提出了一个数据结构的存储规范，提出并比较了几种在ORACLE数据库平台下存储模式的优缺点和选择条件，这为今后此类系统的开发打下了坚实的技术理论基础。 最后对于将数据仓库和数据挖掘的新技术应用于高光谱数据进行了初步的探讨，展望其应用远景，为将来在高光谱数据库系统中集成数据仓库进行了理论探索。
【Abstract】 As the one of newest area of Remote Sensing technology, the hyperspectral remote sensing is used widely and widely，correspondingly the data are richer and richer. With the quantitative remote sensing is more concerned, a strong, powerful, convenient，expandable database system, which stored hyperspectral data and is combined with application function is urgent desired by the related areas. At the same time, the hyperspectral data have their own characteristic. The numbers of wavelength may be varied with the different hyperspectral equipments. One hyperspectral object can be responding to many wavelength data, image data, and many other properties. The quantities of them are great and the relation of them is complexity. And the unique characteristic of hyperspectral data is the united with the image and spectrum. Therefore the spectral database system that we have can not meet the requirements of all these characteristic. Summarizing the results of forerunners, with the research of the status of spectral database now, according to the demands of the hyperspectral data application, this article will first define the concept of Hyperspectral Database in the world. It not only inherits the properties of spectral database system which can store and display the spectrum and other corresponding properties of objects, but also designs the data structure, stores and displays for hyperspectral data. It reflects the characteristic of hyperspectral data from the database area, makes the data structure more compact, combined the image and spectrum, and prepares for the application of data mining. Hyperspectral Database System mainly involves several subsystems, that are hyperspectral image data sub-database, spectral assistant sub-database, hyperspectral data analyzing sub-system, hyperspectral data-warehouse with data mining, and foregrounding system. A Hyperspectral Database System prototype system was completed accompanied with these designs in this article. It includes: Spectral database based in the internet: it comes from the original spectral database system. It is advanced for it is built in the large ORACLE database platform, for it is accompanied by the data analyzing sub-system and for it is published in the internet. It will act as the assistant sub-database for the whole Hyperspectracl Database Systelm; Hyperspectral data analyzing sub-sytem based in the internet: It has basic hyperspectral data analyzing function and is built for expanding the database system applications. It will act on the varieties data of database system; Hyperspectral image data sub-database: it is the soul of the whole Hyperspectral Database system. It will realize the unique characteristic of hyperspectral data, which unites the image and spectrum in the database area. The spectrum can be extract from the image by database. This database, which combines the spatial and spectral information, will play an important role in the future; Foregrounding system: in the times of IT, using internet is the trend. Develop the foregrounding system in the internet will promote the communication of hyperspectral data and exploit the hyperspectal application. This paper will take the concept of Hyperspectral Databse System as the core, state the technology routes, basic frames, special problem in processing, how to solve them and the results in detail and introduce the comparely whole process of building a Hyperspectral Database System prototype. At the same time, in the process of designing, solving problems and programming, this article will summarize and provide the data storing standard rules about data structure. It will also provide comparison of several data storage models under the ORACLE database platform, two of that were accomplished by this system, and state their advantages, shortcomings and choosing condition. This job will make a good technology theory basis for development of this kind system. Finally the article makes an initial discussion of applying the data-warehouse and data mining technology in the hyperspectral data, which makes a theory discussion for integrating the data-warehouse in the Hyperspectral Database System.
(责任编辑：admin) |
