【作者】 刘永山；

【导师】 郝忠孝；

【作者基本信息】 哈尔滨理工大学 ， 测试计量技术及仪器， 2007， 博士

【摘要】 随着空间数据库技术和地理信息系统(GIS)的不断发展和应用,空间推理的基础理论以及相应算法也在不断的创新和发展。目前空间推理主要分为拓扑关系推理、主方向关系推理和距离关系推理等。主方向关系推理是空间推理研究领域的重要分支,在空间推理的研究中有关主方向关系推理的研究越来越引起人们的注意。方向关系模型在空间主方向关系推理的研究中是至关重要的。本文就已存在的方向关系模型进行了研究总结和比较,采用了比较适合空间关系运算的九方向关系模型。人们在空间方向关系的研究中常采用的空间数据类型分为“点”,“线”,“面”。“面”物体在空间数据库中常以最小边界矩形MBR(Minimum Bouding Rectangle)的方式表示。本文以物体的MBR为模型,利用区间代数及矩形代数理论,提出了一种基于MBR的主方向关系与矩形代数关系相结合的推理方法;同时利用方向关系矩阵表示物体MBR之间的方向关系,得到了基于方向关系矩阵的组合推理方法。文中首先针对基于点物体的空间主方向推理进行讨论,在此基础上把点物体的主方向关系推理引深为基于物体的MBR主方向关系推理的问题。结合著名的Balbiani的矩形代数理论,提出了一种基于主方向物体MBR与矩形关系代数相结合的新模型。利用该模型可以很好地将矩形代数良好的计算性质应用于空间方向推理中,简化了方向推理地难度,实现了基于矩形代数的主方向关系一致性检验。对于基于MBR方向关系组合推理的问题,本文使用方向关系矩阵表示物体MBR之间的方向关系。根据提出的一系列方向关系矩阵的性质和方向关系矩阵之间运算的定义、定理,提出了基于矩阵的方向关系之间的组合问题求解方法;针对基于MBR模型求解方向关系组合所存在的问题,提出了两种信息表示模型:深度方向关系矩阵和5×5方向关系矩阵,给出在5×5的方向关系矩阵上方向关系组合的扩展方法。更多还原

【Abstract】 With the development and application of the spatial database technology and geographic information systems (GIS), the basic theories of spatial reasoning and the arithmetic corresponding to them are innovating and developing. At present, spatial reasoning mainly consists of topological relation reasoning, cardinal direction relation reasoning and distance relation reasoning, and so on. As an important division of the spatial reasoning, the cardinal direction relation reasoning has received more and more attentions.The direction relation model is crucially important in the research for the spatial cardinal direction relation reasoning. This paper does some researches and comparisons on the existing direction relation models and uses the nine-direction relation model that is suitable for the operation on spatial relations. The spatial data types which are usually used in the research for the spatial direction relation contains“points”,“lines”and“polygons”.“Polygon”object is usually realized in form of the Minimum Bounding Rectangle (MBR) in the spatial database. This paper takes advantage of interval algebra and rectangle algebra theories. and brings forward a new reasoning method based on combination of the MBR model’s cardinal direction relation and rectangle algebra relation; And provides the composing method based on the direction relation matrix by using the direction relation matrix to denote the direction relations between the objects’MBRs.This paper firstly discusses the spatial cardinal direction reasoning based on the point object, and then develops it into the cardinal direction relation reasoning problem based on the object’s MBR. Combining the famous Balbiani’s rectangle algebra theory, this paper provides a new model based on the combination of the cardinal direction object’s MBR and the rectangle relation algebra. Using the model, the nicer computing character of the rectangle algebra can be well applied in the spatial direction reasoning, which greatly reduced the difficulty of direction relations reasoning, and the consistency verification of the cardinal direction relation based on the rectangle algebra is implemented.As for the direction relations composing problems based on the MBR, this paper uses direction relation matrix to denote the direction relation among the objects’MBRs. According to a series of the definitions put forward on the direction relation matrix’s character and computation, the theorem has provided the method to answer the problem on the composing between direction relations based on the matrix. Aim at the problem about the inconsistent of some composing results from the direction relation model based on the MBR ,this paper construct two information models :depth direction relations matrix and 5×5 direction relations matrix and put forward the method to answer the composing problem using 5×5 direction relation matrix.更多还原