【作者】 姜亚莉；

【导师】 关泽群； 秦昆；

【作者基本信息】 武汉大学 ， 地图学与地理信息系统， 2005， 硕士

【摘要】 图像中包含了大量的信息,本文主要目的是提取图像中结构化、半结构化及非结构化的信息用于决策问题,但大多数技术只能对结构化的问题进行处理,智能决策支持系统正是面向半结构化和非结构化决策问题的,其支持问题的范围包括从纯描述性的非结构化决策问题到常规性的结构化决策问题。智能决策支持系统正是用于研究和解决决策问题中的半结构化及非结构化问题的有效工具。 本文针对二维平面中二值图像目标实体采用启发式的A~*搜索算法进行图像路径的连通性分析,并将智能决策支持系统的相关技术运用于图像中两点间的最优路线规划。 本文将二值图像转换成连通图的形式,采用了RLC的数据结构压缩图像数据,路径的搜索工作都是在连通图的基础上进行的。本文实验引入了符号影射的概念,利用符号推理规则在由图像信息导出的连通图知识结构上进行启发式A~*搜索。符号化的方法使得计算机的功能有质的飞跃,不仅可进行数值计算还可以进行公式推理、符号处理,特别是实际应用中往往希望得到问题的解析模型,但有时往往局限于数学方法及计算机工具发展的程度,而符号化方法就为此带来了便利。符号化的描述方法用于路径规划能真实的、直观的、有效的展现不同实体的空间关系,在分析查询中符号化方法能作为分类和控制的指导。在实验的启发式模块中,A~*搜索和推理发生器协同工作,搜索结果以图形(路线图)及文字(符号规划表)的形式来表示图像的连通性。 空间分析是一门基于地理对象的位置和形态特征的数据分析技术,利用空间分析方法不仅可以查询数据库系统中的各种信息,而且可通过这些信息去揭示事物间更深刻的内在规律和特征。连通性分析作为空间分析的一种广泛地应用于各个领域,如最佳路径搜寻以及其它网络流程分析应用中等。图像中最佳连通性问题即为最佳路径问题,最佳路径问题是一种计算机图形搜索算法,即在出发点和目标点之间找出总代价最低的路径,即尽可能降低算法的时间复杂度和空间复杂度。 本文分析讨论了最短路径的搜索算法,并提出了启发式的路径搜索算法,根据智能决策支持的相关知识结构设计了图像最佳连通性的流程图,对搜索过程进行开发和实现。实验结果表明,启发式的A~*搜索和符号推理规则相结合的方法给图像中两点间最短路径规划提供了一个形象而直观的描述方法,再现了图像实体间的空间关系,并为空间分析和分类提供依据。更多还原

【Abstract】 Images include a large amount of structured and half-structured and unstructured information, usually we can only deal with the structured problem, but IDSS is faced to half-structured and unstructured decision issue, it supports range from unstructured decision issue to structured decision issue.The paper adopts heuristic A* search algorithm used in the connectivity analysis of the route planning between the objects in binary images, and correlative methods of IDSS is used between two points in the images.The paper changes binary images into the connectivity graph firstly, following search processing carried on on the basis of connectivity graph. Experiment introduces the concept of symbolic projection, utilizing symbolic inference rule carry on heuristic searching based on the knowledge structure from image information. The symbolic method made the function of the computer have qualitative leaps, it can used not only in number value and calculate but also in the reasoning of formulae and symbolic processing. Especially often hope to get the analytic model of the question in practical application, but often confine to the degree of the mathematics method and computer tool development sometimes, and the symbolic method brings the facility for this. Adopt the description of symbolic method lies in it can represent different spatial relationship of objects in planning in route truly, ocularrly and validly. And the symbolic method can control in inquiring and analysis as classify and guidance.In the key module of the experiment—heuristic module, A* search process and inference engine operate in tandem. The output of the experiment expresses the connectivity in binary images in the form of route graph and symbolic plan.Spatial analysis is a technology of data analysis based on the geographical objects’s position and shape characteristics, utilizing the spatial analytical method not only can inquire about various kinds of information in the database system, and can announce deeper inherent rule and characteristic among the information in geographical objects. Connectivity analysis, as one of the spatial analysis, is widely applied in many fields, such as the shortest path problem and the procedure analysis of the whole network and so on. The shortest path problem in images is a kind of computer graph searching algorithms. The optimum algorithm of the route should finish searching for the route of the minimum cost. Namely reduce time complexity of the algorithm and space complexity as much as possible.The paper designs the flow of the best connectivity in images on the basis of the correlative knowledge structure in IDSS. The experiment show that the result of the shortest route searching using the method of heuristic A* search algorithm and symbolic reasoning is better than of no heuristic search algorithm. And the method reproduces spatial relationship of the objects in images ocularrly, offers basis on which the spatial analysis classify.更多还原