【作者】 申世群；

【导师】 刘大有；

【作者基本信息】 吉林大学 ， 计算机应用技术， 2011， 博士

【摘要】 空间推理是指利用空间理论和人工智能AI(Artificial Intelligence)对空间对象进行表示和推理。近年来,空间推理已成为十分活跃的研究领域,在空间演算易处理性,多维空间演算,空间知识管理以及定性、不完备空间信息推理等方面取得了重要进展。随着空间信息技术的发展,空间推理逐渐被用于地理信息系统、空间知识发现、时空数据库、图像数据库、自然语言理解和多媒体数据库等领域。其中地理信息系统是空间推理的最重要的应用领域之一。近年来,伴随着信息获取技术的进步,在地理信息系统( GIS)和遥感图像处理等领域中的空间数据,呈现出爆炸式增长的趋势。那么如何更加有效地利用这些空间数据、如何在海量空间数据中快速检索出人们所需要的信息成为目前空间数据管理的瓶颈之一,因此GIS的交互问题越来越重要,在这方面有很多的问题亟需研究解决,如以矢量或栅格方式存储着几十亿字节数据的GIS系统现在还不能提供直观的、面向常识的人机交互功能,诸如GIS系统不支持从大量数据中抽取出定性空间信息等。要想解决这些问题,就要对定性空间推理进行深入研究,并将定性空间表示、推理和空间相似性等研究结果应用到GIS中去,使GIS能满足人们日益提高的检索需求。本文在分析现有工作的基础上,围绕定性空间推理在空间数据检索中的应用开展研究,主要研究结果如下:1.研究了定性空间推理的相关概念和基本方法,对定性空间表示和空间数据检索进行了总结,通过对比分析重点研究了九交集拓扑关系模型和方向关系矩阵模型,提出结合这两种定性模型进行空间数据检索的方法,能对空间草图进行有效地表示和检索。总结出空间推理研究的基本方法主要有公理化方法、几何约束满足方法、代数方法和基于模型的推理方法。研究了判定某种定性空间关系形式化表达能力的标准,目前定性空间表示研究主要集中在空间拓扑关系表示、空间方向关系表示、空间距离关系表示、定性形状表示、空间邻近关系以及结合多种空间关系的定性表示方法,在这些表示方法中最重要的是拓扑关系和方向关系的表示,其中应用较广泛的拓扑关系模型主要有RCC模型和九交集模型,而应用最广泛的方向关系模型是方向关系矩阵模型。空间数据检索是近几年的研究热点,随着人们获得的空间数据不断增多,迫切需要提高空间数据检索的效率,增加更多、直观、符合人们认知的检索方法。我们介绍了空间数据检索发展的阶段及未来的发展方向,现阶段对基于草图的空间数据检索的研究越来越多,有必要对该方法进行深入研究。2.基于结合九交集拓扑关系模型和深度方向关系矩阵模型,研究了基于草图的空间数据检索方法。在基于草图的空间数据检索中,用户可用鼠标或手在触摸屏上绘制将要检索的空间场景,如建筑物、桥梁、河流和山川等形状与空间相对位置信息,并可给出已知的标注,形成草图。草图包含了较明确、详细的信息,包括对象间空间拓扑关系、方向关系和已知对象的标注等,可作为检索条件提交给GIS,系统对草图中的对象本身及对象间的拓扑和方向关系进行提取,并将提取的特征记录在相应的特征关系表中,然后通过空间关系匹配算法检索到符合要求的空间数据并进行显示,实现基于草图的空间数据检索。近年来,基于草图的空间数据检索得到了重视和研究。自1996年以来,Egenhofer、Blaser等人都对基于草图的空间数据检索进行了研究,相继给出了基于草图的空间数据检索的设计原则,草图的表示及检索处理过程,但以往的研究中大部分都围绕着区域对象展开,没有考虑所有类型的空间对象,并且很多都没有给出系统原型,我们将九交集拓扑模型和深度方向矩阵引入空间数据检索,给出了一种基于草图的空间数据检索方法,该方法支持地理数据库中所有的数据类型。我们具体研究了空间草图中拓扑关系和方向关系的提取及保存方法,并将基于草图的空间数据检索问题转化为约束满足问题,并根据约束满足问题的前项检查算法的思想,针对空间数据检索的具体问题,给出了一个基于草图的空间数据检索算法SBSDQ-FC(),应用标注、定义域动态排序和空间邻近关系等方法对SBSDQ-FC()算法进行了改进,提高了草图检索算法的检索效率。并给出了算法的复杂度,通过实验对算法进行了分析验证。3.空间相似性及其在空间数据检索中的应用研究研究了空间相似性的概念和相关处理方法,综述了空间相似性的国内外研究现状。重点研究了拓扑关系和方向关系相似性的定义和计算方法。根据1996年Bruns和Egenhofer给出的任意两个区域之间的拓扑关系概念邻域图,通过计算任意两个拓扑关系的距离得到两个区域之间拓扑关系概念邻域的差异矩阵。根据概念邻域图和差异矩阵给出了面与面之间拓扑关系相似性的计算方法。同时将这个方法推广到其它对象间拓扑关系相似性的计算。对于方向关系相似性计算,主要从主方向关系模型出发,研究了基于主方向关系模型的方向关系相似性计算方法。最后将空间相似性研究结果应用于基于草图的空间数据检索,使检索方式更直观,更易于理解。4. GIS环境下基于草图的空间数据检索系统的设计与实现为验证本文提出的结合拓扑关系和方向关系的草图检索方法,我们应用C#和MapInfo建立了一个基于草图的空间数据检索原型系统,验证了我们所提出的方法的可行性,同时文中也分析了这种方法的不足,指出了下一步要做的工作。国内空间推理领域关于定性空间推理及其应用研究方兴未艾,本文以上的研究结果丰富了定性空间推理及其应用技术,期望对该领域的发展有一定的借鉴和参考。更多还原

【Abstract】 Spatial reasoning is the representation and reasoning of the space object used spatial theory and artificial intelligence. In recent years, the study of spatial reasoning has become a very active research field and has made great progress in qualitative spatial reasoning, spatial knowledge management etc. With the development of information technology of space, spatial reasoning now is being used in geographic information systems, spatial knowledge discovery, image database, multimedia database etc. Geographic Information System is one of the most important fields of spatial reasoning applications.In recent years, with the development of information technology of space, the spatial data is now showing a trend of explosive growth in geographic information system (GIS) and remote sensing image processing and other fields. So how to make effective use of these spatial data becomes one of the bottlenecks of spatial data management. The interaction issues of GIS becomes more and more important, there are many problems need to be solve in this field, such as there are billions of bytes of data stored in the GIS database by vector or raster mode, but the GIS system can not provide an intuitive method for interaction. For example, a user may want to extract some qualitative data from a large number of spatial data or submit a qualitative query to the GIS system, but now the GIS system can not provide such function. To solve these problems, we should study qualitative spatial reasoning and apply the research results of qualitative spatial reasoning, qualitative spatial representation and spatial similarity to the GIS system, so that GIS can fully meet the increasing needs. In this paper, we first give a analysis of existing work, then study on qualitative spatial reasoning and the application of qualitative spatial reasoning in spatial data query and other fields.The main results are as follows:1.Studied the concepts, research contents and the basic methods of spatial reasoning .Summaried the qualitative spatial representation and spatial data query methods, we focused on the 9-intersection model of topological relation and the Deep-Direction-Relation Matrix model of the direction relation.We will use the combination of these two qualitative models in spatial data query by sketch.Summarized the basic methods of spatial reasoning research.The methods includes the axiomatic method, the geometric constraint satisfaction methods, algebraic methods and model-based reasoning. The qualitative spatial representation research focuses on spatial topology relation representation, spatial direction relations representation, spatial distance relation representation, the qualitative shape representation and the qualitative representation method of multi spatial relation etc. the most important representation is the representation of the topological and direction relations. The 9-intersection model and RCC model are the most important topological models; The Deep-Direction-Relation Matrix model is the most widely used direction model. Spatial data query is a research hotspot in recent years, with the growing of the number of spatial data, people need to improve the efficiency of spatial data query, and increase more intuitive way to meet people’s awareness of query. We introduced the stage of development of spatial data query, and the future of the spatial data query. We’re seeing an ever increasing number of people interested in studying of the spatial data query by sketch now. It is worthy of in-depth research and concerns.2. Based on the combination of the 9-intersection model and the deep direction relation Matrix model, we studied the sketch-based spatial data query methods.Spatial database contains a great deal of topological and directional semantics. But traditional spatial data query methods didn’t make good use of these high level semantics. To overcome this conceptual gap, this paper proposes a spatial data query method based on sketch using 9-intersection model and Deep-Direction-Relation Matrix. This method integrates direction relations and topological relations and can handle all data types in geographical databases. This thesis outlines an algorithm based on the solutions of Binary CSP. A prototype has been developed to experiment the method this thesis proposed.3. Spatial similarity theory and its application in spatial data quey by sketch.This thesis summarized the spatial similarity theory and its application in spatial data query.We also study the concept of the spatial similarity and its research content. We outline the present status of research on spatial similarity theory both in China and abroad. Studied the definition of the spatial similarity of the topological relation and its calculation methods.We also studied the definition and calculation methods of the spatial similarity of the direction relation.the application of spatial similarity in spatial data query has also been studied.4. The design and implementation of the spatial data query by sketch system in GIS environment. To verify the proposed method of spatial data query based on sketch using 9-intersection model and Deep-Direction-Relation Matrix, we developed a prototype system used the C# and MapInfo, introduced the design, query processes and realization of the prototype system.We have done some related experiments on the prototype system, we demonstrated the feasibility of the proposed method, also pointed out the shortcomings and the application prospects of the method we proposed.More and more people began to pay close attention to the study of qualitative spatial reasoning and its application, this work is expected to promote the development of the qualitative spatial reasoning and its application. In this paper, the spatial data query method based on sketch using 9-intersection model and Deep-Direction-Relation Matrix proved to have high theoretical and practical value.更多还原