【作者】 陈旭；

【导师】 马宗民；

【作者基本信息】 东北大学 ， 计算机应用技术， 2017， 博士

【摘要】 现实世界中许多应用领域,诸如机器视觉、数字影像和地理信息系统等都与时间和空间信息紧密关联,如何有效地管理时空信息显得愈发重要。时空数据模型作为时空数据管理的基础,对时空数据管理中后续的存储、查询以及分析等起到了关键性作用。因此,如何有效建模时空数据成为一个重要的研究课题。统一建模语言UML(Unified Modeling Language)作为面向对象开发领域的行业标准,已经在软件工程和数据建模等许多领域得到了广泛的应用。UML类图丰富的表达能力为复杂的时空语义信息建模奠定了基础。此外,XML(eXtensible Markup Language)是Web信息交换和传输的事实标准。随着时空信息在Web领域的广泛应用和传播,XML的时空扩展已成为时空数据建模的重要形式。在现实世界中,信息通常是不精确的。大量的时空数据同样存在模糊性,而现有的时空数据模型缺乏描述和表达模糊时空对象内在机制和语义关系的能力。因此,为了有效地表示和处理带有模糊性的时空数据及时空事件,本文重点研究了基于UML和XML的模糊时空建模与转换方法。首先通过研究模糊时空数据语义,给出了模糊时空数据的形式化定义。在此基础上,根据模糊时空数据在概念和逻辑方面的建模需求,对UML和XML进行了模糊时空扩展,提出了模糊时空UML概念数据模型和模糊时空XML逻辑数据模型,并通过实例验证了模型的可用性和有效性。最后,实现了模糊时空UML模型与模糊时空XML模型之间的相互转换,给出了相应的转换规则和实例。本文的主要贡献和具体的创新性研究成果包括以下几个方面:(1)通过研究和分析时空数据的模糊特性和语义关系,给出了模糊时空数据模型的形式化定义。首先,给出了模糊时间点和区间、模糊空间点、线和区域的形式化定义;然后,深入分析讨论了模糊时间元素、空间元素的方向、度量和拓扑关系,给出了相应的形式化表示形式;同时,针对动态时空现象的表示问题,提出了模糊时空事件-过程的表示方法;最后,给出了模糊时空数据模型的形式化定义。(2)提出了一种模糊时空UML概念数据模型。通过扩展UML模型,研究了基于UML的模糊时空类、模糊时空数据、模糊时空事件以及模糊时空概化、聚集、关联的语义表示方法,进而提出一种模糊时空UML概念数据模型,最后以藤原效应为例验证了模型的可用性。(3)提出了一种模糊时空XML逻辑数据模型。首先,给出了模糊时空数据在XML文档中的表示形式;然后,在XML树形结构的基础上给出了模糊时空XML数据树;在此基础上,对XMLSchema进行了模糊化扩展研究,提出了模糊时空类、模糊时空数据、模糊时空事件及模糊时空语义关系的XML Schema表示方法;最后,以气象现象为例建立了基于XML的模糊时空数据模型。(4)提出了模糊时空UML模型与模糊时空XML模型的相互转换方法。首先,模糊时空UML数据模型能在概念层表示丰富的时空语义,但是在模糊时空数据查询以及存储方面存在一定的局限性,为此提出了一种从模糊时空UML数据模型到模糊时空XML数据模型的转换方法,给出了相应的转换规则;其次,考虑到XML在语义表达上不够丰富和直观,提出了从模糊时空XML数据模型到模糊时空UML数据模型的转换方法,给出了相应的转换规则;最后,结合转换实例验证了方法的可用性。更多还原

【Abstract】 In the real-world,many application areas such as computer vision,digital imaging and GIS,are closely related to temporal and spatial information,effective management of spatiotemporal information thus becomes more and more important.The spatiotemporal data model,as the basis of spatiotemporal data management,plays a key role for the follow-up storage,querying,and analysis in spatiotemporal data management.Therefore,how to effectively model spatiotemporal data has become an important research topic.UML(Unified Modeling Language),as an industry standard for object-oriented development,has been widely used in many fields such as software engineering and data modeling.The rich expressive power of UML class diagram lays a foundation for spatiotemporal data modeling.In addition,XML(eXtensible Markup Language)is the defacto standard for Web information exchange and transmission.With the widespread application and dissemination of spatiotemporal information on the Web,XML-based spatiotemporal extension research has also become an important form of spatiotemporal data modeling.In the real-world,information is often imprecise.A large number of spatiotemporal data are also ambiguous,and existing spatiotemporal data model lacks the ability to describe and express the inherent mechanism and semantic relations of fuzzy spatiotemporal objects.Therefore,to well represent and process spatiotemporal data and spatiotemporal events of fuzziness,this dissertation focuses on fuzzy spatiotemporal modeling and transformation based on UML and XML.Firstly,the formal definition of fuzzy spatiotemporal data is given by studying the semantics of fuzzy spatiotemporal data.Based on the conceptual and logical modelling requirements of fuzzy spatiotemporal data,the fuzzy spatiotemporal extensions of UML and XML are proposed.The fuzzy spatiotemporal conceptual data model of UML and the fuzzy spatiotemporal logical data model of XML are proposed.The usability and effectiveness of the models are proved by examples.Finally,the transformation between the fuzzy spatiotemporal UML model and the fuzzy spatiotemporal XML model is implemented,and the corresponding transformation rules and examples are given.The main contributions and the specific innovative research results of this dissertation include:(1)By studying and analyzing the fuzzy characteristics and semantic relations of spatiotemporal data,an abstract formal definition of fuzzy spatiotemporal data model is given.Firstly,the formal definitions of fuzzy time points,fuzzy time interval,fuzzy space points,lines and regions are given.Then,the direction,metric and topology relations between fuzzy temporal and spatial elements are deeply analyzed and discussed,and the corresponding formal representations are given.At the same time,the representation method of fuzzy spatiotemporal event-process is proposed for the representation of dynamic spatiotemporal phenomena.Finally,an abstract formal definition of fuzzy spatiotemporal data model is given.(2)A fuzzy spatiotemporal conceptual data model of UML is proposed.By extending the UML model,a UML-based semantic representation method of fuzzy spatiotemporal classes,fuzzy spatiotemporal data fuzzy spatiotemporal events,fuzzy spatiotemporal generalization,fuzzy spatiotemporal aggregation,and fuzzy spatiotemporal association are studied.A fuzzy spatiotemporal UML conceptual data model is thus proposed.Finally,the usability of the model is verified by the Fujiwara effect as an example.(3)A fuzzy spatiotemporal XML logical data model is proposed.Firstly,the representation form of fuzzy spatiotemporal data in XML documents is given.Then,the structure of fuzzy spatiotemporal XML data tree is given based on XML tree-shape representation.On this basis,the fuzzy extension of XML Schema is studied.The representation methods for XML Schema of fuzzy spatiotemporal classes,fuzzy spatiotemporal data,fuzzy spatiotemporal events,and fuzzy spatiotemporal semantic relations are proposed.Finally,as an example,a XML-based fuzzy spatiotemporal data model for meteorological phenomena is established.(4)A transformation method between fuzzy spatiotemporal UML model and fuzzy spatiotemporal XML model is proposed.Firstly,the fuzzy spatiotemporal UML data model can represent rich spatial and temporal semantics in the concept layer,while has limitations in the storage and querying of fuzzy spatiotemporal data.For this reason,a transformation method from fuzzy spatiotemporal UML data model to fuzzy spatiotemporal XML data model is proposed.Secondly,considering XML is not rich and intuitive in semantic representation,a transformation method from fuzzy spatiotemporal XML data model to the fuzzy spatiotemporal UML data model is proposed,the corresponding transformation rules is given.Finally,in combination with a transformation example,the usability of the proposed method is verified.更多还原