【作者】 吴元洪；

【导师】 郭平；

【作者基本信息】 重庆大学 ， 计算机应用技术， 2003， 硕士

【摘要】 空间索引技术在计算机辅助设计与制造（CAD/CAM）、地理信息系统（GIS）、图像处理（image processing ）、虚拟现实语言设计（VRML）、数字地球（digital earth）等诸多领域均具有十分重要的研究意义，它能为GIS中图形图像的存取处理提供技术支持，为空间关系的分析处理提供理论依据，同时为空间数据库的设计者在采用的数据结构方面提供有益的参考。本论文首先对空间数据库的索引进行探讨，并给出了空间索引结构的发展演化图，然后介绍了目前国内外GIS的空间连接过程中的主流索引结构，最后引入了字符来表达空间对象。在大比例尺空间及日常生活中,人们普遍使用定性描述理解、分析和对空间环境下结论。定性描述可以处理非精确数据，简化描述和推理过程。当前地理信息系统空间关系描述模型和表示大多是定性的，而自然语言描述中往往采用定性的方法，空间信息的定量处理方式明显与人们对空间关系下结论的方式不同，不符合人们的空间认知结构。本文主要介绍了利用字符对二维空间的空间对象进行定性空间分析。本文着重讨论了GIS中二维简单空间区域对象的空间关系， GIS空间数据索引方式，以及空间数据索引的应用。本文介绍了一种基于字符的空间对象索引方式及其在表达空间对象的空间关系分析上的应用。在空间连接处理的算法上采用了经典的R树作为索引，采用批生成算法生成R树。为改善生成效率，对空间对象的MBR按矩形中心点进行Hilbert排列码排序。实践证明，这种方式取得了最好的查询性能。在求精时，摒弃了传统的复杂的几何计算，本文提出了一种基于字符的查找模式，将二维的空间对象转化为一维字符串进行处理，并利用启发式搜索算法将二维空间上的无序查找转化为有序查找，从而利用折半查找法，大大提高了查询效率。实践证明，字符串在表达空间对象的方向关系上较为有效。本文基于对空间对象的投影，生成两个一维方向上的字符串，然后通过分析字符串，利用空间方向关系矩阵来进一步确定方向关系，进而得出空间对象的相对方向关系。更多还原

【Abstract】 The Spatial Data Index Technology is of significance in researching into CAD/CAM, GIS, Image process, VRML, digital earth and many other fields. It can provide technological support for accessing images in GIS, offer theoretical basis for analyzing and managing space relation, and give spatial database designers helpful references for adopting data structure.This thesis discusses spatial data index first, shows the history of spatial index structure, introduces the current domestic and foreign main index structure in GIS’s spatial joining, and then makes use of character to express spatial object. Qualitative description is often used to understand, analyze and conclude about spatial environment in large scale space and everyday life. Qualitative description can deal with unaccurate data and simplify the description and reasoning process. The representation model of spatial relation and its expression are mostly qualitative in the current GIS, and natural language description often adopts the qualitative method. The quantitative method of spatial data is apparently different from the method of people’s concluding about spatial relation and not in accordance with spatial cognition. This thesis is intended for the introduction to the qualitative spatial analysis of spatial object in two-dimensional space by using character.This thesis mainly discusses spatial relation of simple area object in two-dimensional space, the methods of spatial data index in GIS and the application of spatial data index. This thesis introduces a method of spatial data index based on character and its application to spatial relation analysis of spatial object. The classical R-tree as index is adopted in spatial joining. R-tree is performed by bulk loading. In order to perfect performance efficiency, the MBR of spatial object is ordered by the Hilbert ordering code of the center of spatial object’s MBR. Practice shows that this method gets best query performance. When refining the candidates, the author gets rid of traditional complicated geometric calculating. This thesis sets forward a new query model based on character. This model changes two-dimensional spatial object into one-dimensional string to process, and therefore turns two-dimensional disorder query into order query by using heuristic querying algorithm so that bisearch is used to improve query efficiency greatly. It is proved that string is relatively effective on expressing direction relations of spatial object. Based on the projection of spatial object,<WP=5>two one-dimensional strings are produced. Analyzing these strings and using spatial direction relation matrices, we could further revise the direction relation. As a result the relative direction relation of spatial object is obtained.更多还原