一种新的海量遥感瓦片影像数据存储检索策略

发布时间：2018-03-29 04:30

本文选题：瓦片影像　切入点：嵌入式数据库　出处：《电子科技大学》2013年硕士论文

【摘要】：遥感影像作为数字地球建设计划的基础地理信息数据，其高效存储和快速检索是遥感影像信息提取和共享的重要手段。近几年，，随着高分辨率对地观测技术的发展，遥感影像数据量呈几何级增加，对于海量的影像数据如何才能有序、高效地存储，组织与管理，并形成统一的存储组织规范，从而实现遥感数据的共享与快速发布服务，已经成为空间信息科学领域急需解决的问题。通过对国内外典型的海量遥感数据存储管理系统的研究，不难发现采用多分辨率金字塔和影像分块技术，是实现海量遥感影像信息的实时显示和高速服务的有效途径。然而采用影像金字塔技术组织遥感影像数据本质上采用的是牺牲存储空间的方式，因此必然会导致遥感影像的数据量大幅增加，使得本来已经呈现海量状态的遥感影像数据再次呈倍数增加。由此可知海量遥感瓦片数据的高效存储及管理具有重要意义，本研究正是基于此热点问题，通过分析当前国内外比较知名的海量遥感瓦片影像存储管理系统，结合瓦片数据的空间特性，提出一种基于轻型嵌入式数据库和文件系统结合的海量遥感瓦片影像存储方案，该方案采用分布式文件系统存储瓦片影像实体数据，瓦片的索引数据或元数据存放到各节点树形存储结构下的嵌入式数据库文件，瓦片影像数据与元数据都实现了分布式存储的功能，从而解决了集中式元数据存储查找的瓶颈问题。具体的分布式策略借鉴哈希取模算法，但是考虑到瓦片影像数据的安全性和分布式存储系统的可用性，本研究提出并设计了一种基于哈希模的冗余策略-哈希主副配号法，并且定性的分析了采用哈希冗余策略设计的分布式文件系统的可靠性，可扩展性和负载均衡性等特点；本研究的存储策略决定了检索方式即能支持直接寻址的精确查找又能支持基于数据库SQL技术的模糊查找，文中详细的阐述了检索原理、实现过程和核心代码，并且进行了大量的实验测试，通过与传统的基于大型商用数据库的集中式查找技术作比较，表明本研究的检索效率具有明显的优势，能够满足海量数据源查找请求。文章最后总结了本研究的主要贡献和创新点，并提出了下一步的研究计划。
[Abstract]:Remote sensing image as the basic geographic information data of the digital earth construction plan, its efficient storage and fast retrieval is an important means of remote sensing image information extraction and sharing. In recent years, with the development of high-resolution earth observation technology, The amount of remote sensing image data is increasing in geometric order. How to store, organize and manage the massive image data in an orderly and efficient manner, and form a unified storage organization standard, so as to realize the sharing and rapid release of remote sensing data. It has become an urgent problem in the field of spatial information science. Through the research on the typical massive remote sensing data storage and management system at home and abroad, it is not difficult to find that the multi-resolution pyramid and image block technology are used. It is an effective way to realize the real-time display and high speed service of mass remote sensing image information. However, using image pyramid technology to organize remote sensing image data is essentially a way of sacrificing storage space. Therefore, it will inevitably lead to a large increase in the amount of remote sensing image data, which makes the remote sensing image data which has already presented a massive state to be multiplied again. It can be seen that the efficient storage and management of massive remote sensing tile data is of great significance. This study is based on this hot issue, through the analysis of the domestic and foreign well-known mass remote sensing tile image storage and management system. Based on the spatial characteristics of tile data, a massive remote sensing tile image storage scheme based on the combination of lightweight embedded database and file system is proposed. The scheme uses distributed file system to store tile image entity data. The index data or metadata of tile is stored in the embedded database file under tree structure of each node. Both tile image data and metadata realize the function of distributed storage. Therefore, the bottleneck problem of centralized metadata storage and search is solved. The specific distributed strategy uses hash extraction algorithm for reference, but considering the security of tile image data and the availability of distributed storage system, In this paper, we propose and design a kind of redundant strategy based on hash module, which is called hashing master and sideload, and qualitatively analyze the reliability, extensibility and load balance of distributed file system designed by hash redundancy strategy. The storage strategy of this study determines that the retrieval method can not only support the accurate search of direct addressing, but also can support the fuzzy search based on database SQL technology. In this paper, the retrieval principle, implementation process and core code are described in detail. By comparing with the traditional centralized search technology based on large commercial database, the retrieval efficiency of this study has obvious advantages and can satisfy the mass data source search request. Finally, the paper summarizes the main contributions and innovations of this study, and puts forward the next research plan.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208

【参考文献】