摘要本论文结模式识别技术,在基于本体理论的基础上研.docVIP

摘要：本论文结合模式识别技术，在基于本体理论的基础上研究并实现强对流天气推理系统。该系统通过模式识别技术对图像数据进行预处理和特征提取，从中获取有用的特征信息，并结合气象专业知识，对特征信息进行分析，抽象出符合一定气象意义的知识，建立领域本体库，在此基础上研究并设计强对流领域推理规则，进而建立起强对流天气推理系统，以提高强对流气象预报领域的计算机化、快速化和实时化。 本文主要工作针对强对流预报领域应用场景设计并实现了一个基于本体理论的知识表达的推理原型系统。 首先，本论文参考大量气象文献和在气象领域专家的参与下，利用protégé 本体开发工具构建强对流领域本体库，用以描述该领域中的知识概念和概念间的约束和联系。 其次， 详细描述了基于强对流本体库的SWRL规则的建立过程，从分析待解决问题到推理流程的描述，再到swrl规则的Atom的建立和Imp的建立，直到最后完成了完整的规则库的开发，详实的介绍了每一步的开发过程。 再次，本文采用Jess引擎来作为该系统的推理引擎。Jess引擎是根据从已建立的强对流本体库和规则库转换过来的事实和规则进行推理的，因为Jess引擎不能直接推理以OWL本体语言和SWRL表示的知识与规则，所以需要SWRL解析器将为Jess引擎的事实库和规则库做相应的格式转换，将以OWL本体语言和SWRL表示的知识与规则转换为Jess引擎可接受的格式。 最后，本文在强对流领域本体库和SWRL解析器的基础上，构建相应的推理系统，完成推理预测要求，并以实验验证推理系统的正确性。 关键词：本体；SWRL；JESS；强对流；owl；RDF；pretege； Abstract: This paper presents a prototype system, which has based on research in ontology. This system has built an ontology library on the strong convection in meteorology, and SWRL rules in this domain. And then semantic reasoning on the domain was implemented by applying with Jess. This system will computerize the process of weather forecast and improve the forecast efficiency. The main work as follows: 1. After reading many weather references and with help of weather professors, this paper has designed the strong convection ontology database to describe the restrictions and relations of the weather conception using protégé 3.3 beta version 2. This paper has built SWRL rule to reasoning the unknown result in detail, which includes the development of Atom and Imp. 3. After Analyzing and introducing the theory of SWRL parsing machine, translating the SWRL rule format to JESS format, the ontology to JESS fact. Based on ontology database, SWRL rule and Parsing machine, this paper has run the reasoning system and got the results using jess. 4. With mode identify and ontology technology, the paper has present the prototype system of weather reasoning. Keyword: ontology; SWRL; OWL; pretege；RDF; JESS; strong convection TOC \o 1-3 \h \z \u HYPERLINK \l _Toc195938667 第一章 绪论 PAGEREF _Toc