【作者】 贾彪；

【导师】 贺春贵；

【作者基本信息】 甘肃农业大学 ， 农业昆虫与害虫防治， 2007， 硕士

【摘要】 本研究综述了国内外植保专家系统的研究成果,以贺春贵教授所主编“苜蓿病虫草鼠害防治”(2004)一书为基本素材,应用软件工程和专家系统技术,研究开发了苜蓿病虫草鼠害防治专家系统模块应用框架,建立了基于单机和网络运行的苜蓿病虫草鼠害防治专家系统。该系统由系统管理模块等八大模块构成,对苜蓿病、虫、草、鼠害进行识别、查询、防治技术以及网络交互,初步实现了苜蓿有害生物综合治理和科学防治的专家决策体系。苜蓿病虫草鼠害防治专家系统开发平台.NET Framework2.0,由软通动力信息技术有限公司北京总部(http://www.isoftstone.com)协助提供。该平台运用微软智能客户端技术(Smart Client)开发完成的,它既有Windows程序的强大用户界面,又可以享受Web应用升级和更新的便捷。可以说是充分利用了Windows和Web技术的优势。其每个模块的构建都是通过采用“浏览器/Web服务器/数据库系统”3层网络结构模型实现的。平台面向对象的程序语言是Microsoft Visual C#,辅助以InterDev,Asp,HTML语言,Dream Weaver网页制作工具,并以Microsoft SOL Server2005作为后台数据库开发。在此平台下开发出的苜蓿病虫草害专家系统采用了Windows操作界面,菜单驱动,易学易用,既可以单机运行,又可以在网络环境下运行,支持多用户同时访问。针对不同用户拥有不同的开发环境。用分布式技术、构件技术、Web技术、面向对象技术、这样的分层结构可以使本系统方便的进行升级和完善。开发人员接到用户的反馈后,可以不断修改和完善系统。本系统涉及了苜蓿主要虫害50种、病害40多种、杂草30多种、鼠害4种作为研究对象,并利用面向对象的知识表示方法来进行苜蓿病虫草鼠害知识的计算机信息表达,使知识获取、知识表示、和利用有机的结合起来,在一定程度上克服了知识获取的“瓶颈”问题。采用人工的系统学习机制来保证知识表达的准确性和实用性,通过专家管理界面进行知识的补充、删除和修改。由于苜蓿病虫草鼠害防治专家系统的知识类型复杂,因此,所收集的知识需要经过整理抽取,使知识和信息规范化,才能符合计算机数据信息推理和检索的处理要求,有利于知识的表达和利用。本系统按照病虫草鼠危害的部位进行对象分类,即以苜蓿为父类,整株、根、茎、叶、芽、花、果实为子类进行知识抽取和整理,分别建立苜蓿病虫草鼠害知识库和苜蓿病虫草鼠害关键词知识库,充分利用文字、图片等信息进行知识的表达。本系统在推理策略上选择数据驱动和目标驱动相结合,即采用正反向混合推理的模式,系统操作简单,方便实用,有效的解决了病虫草鼠害“并发症”的诊断问题。根据推理过程中界面知识表达的完整性和用户不同知识水平的需求程度,苜蓿病虫草鼠害专家系统诊断方式分为检索式诊断模式和关键词诊断模式。检索式诊断模式是一种由系统程序引导的判断模式,用户根据界面提示病虫草鼠危害部位和典型症状,利用面向对象的推理策略,经过正反向推理得出具有这些症状的病虫草鼠害中文名、英文名、拉丁学名等,并给出详细信息;关键词式诊断方式是根据用户选择的关键词或输入的关键词进行匹配,经过正反推理,得出具有这些苜蓿症状的病虫草鼠害名称并给出详细的防治方法和信息。更多还原

【Abstract】 In this thesis plant protection expert research and development home and abroad was summarized, and an expert system for pest control in alfalfa (abr. PestCinA) was set up. The expert system was based on PC and .NET Framework2.0 platform. This expert system consisted of eight modules including management module etc. And all the modules can provide identification, enquiry, prevention and web interaction for pest control in alfalfa. It came out a good expert decision-making system for alfalfa pest control for users. All the special knowledge was from the book‘The Control of Pest in Alfalfa’edited by Prof. He Chun-Gui.in 2004.The .NET Framework2.0 platform was provided concurrently by Beijing headquarters of ISOFTSTONE drive information technology limited company (http://www.isoftstone. com). The platform was exploited and completed by using Microsoft intelligent client port technology. It not only concluded strong user interface of windows procedure, but also enjoyed web applied upgrade and prompt update. It used dominance of windows and web technology. Every module’s form is completed by using (browser/web/database system) three-floor net frame model. Microsoft Visual C# and SQLServer2005 is used for the back-end database, and InterDev, Asp and HTML language are also used in the developing process. On this platform, the expert system for pest control in alfalfa used windows interface, menu driver, so it is easy to learn and use, it also can be used in both PC and Internet, support multi-users simultaneously to browse through Internet.There are different exploit environment to different user. Used distributed technology, component technology, web technology and faced objected technology, this kind of layer framework can made the system upgrade and consummate more easily, Exploit personnel can modify and consummate system constantly when received user’s feedback information.Information of the expert system mainly roots invalid date and experts’experience on basis of which, 50 kinds of insect pests, 40 kinds of diseases, 30 kinds of seeds and 4 kinds of mice with the production of alfalfa have been chosen as the subject for research in the system .An object-oriented information representation solves the problem of the “bottleneck”of the expert system on a certain extent, and its’information database can be appended deleted and modified by module.The type of pest control in alfalfa information is complicated which includes qualitative information, quantificational information, fixed information and uncertain information. The expression of information is formation of facts and connections of pest control in alfalfa, i.e. the collection of the date operation on the computer. Also it must be suitable for the expression and utilization of information. Access database and access key word database of pest control in alfalfa have been build up according to the catalogue of the disserve part, namely the whole plant, root, bole and branch, lamina, bud, blossom, fruits, in addition, lots of pictures are used in the system as well.In view of the complexity of the information and the failure to provide reliable information by the users, the PestCinA was set up in favor of combination method of data drive and aim drive, namely the positive and negative commix inference model was more suitable for the practical use of the expert system. According to requirement of the integrity of the expression of information interface and the various levels of users, the diagnosis mode of pest control in alfalfa expert system can be used in two ways: the mode of diagnosis and that of the key word mode for the search of information. The search model of diagnosis was a concluded way by the guidance of computer system. Users can choose disserve part and typical symptom according to the users’guide. Inference engine may use object-oriented strategy to give the name of pest control in alfalfa. The model of key word diagnosis is that the user can chooses or inputs key word to match with normative key word, and the inference engine can also give the name of pest and its control measures in alfalfa that is provided with the same symptom.更多还原