|
Apple cultivation is one of the six pillar industries in Shaanxi, and the main factors affect the yield and quality of apple is an apple pest apple pest and disease diagnosis, how the use of modern information technology to provide a quick way for the timely prevention is an urgent need to study the problem. This article for apple pests and disease diagnosis, prevention and governance requirements analyzes from Apple's disease characteristics, mainly to study the the apple disease diagnosis knowledge acquisition and reasoning method, based on similarity and disease diagnosis expert system based on BP neural network model to build the knowledge base of apple diseases and diagnostic platform and intelligent system platform. The main contents and conclusions are as follows: (1) a comprehensive and detailed analysis of Apple's disease pathogenic factor, impact factor, disease type, disease type of knowledge, the apple disease diagnostic parameters (incidence period, location, the pathologies color, symptoms shape, disease) and disease knowledge representation method for dynamic encoding and automatic generation of disease diagnosis the parameters coding and diagnostic rules encoding and stored in the knowledge base, and analyze the storage structure of the knowledge base, to build a relational database-based Apple Disease Knowledge Base . (2) study disease diagnosis model based on similarity. Definition of the concept of similarity apple disease diagnosis, given Apple's disease similarity judgment methods and standards, using the rough set theory method to determine the the disease rules diagnosis parameter similarity diagnosis model the weight of the same diagnostic parameters in different diseases parameter weights the same conditions, to construct a similarity model and its algorithm platform, and experimental results show that the three groups of test samples, based on the the similarity model diagnostic accuracy rate were 91.70%, 78.40%, 33.30%. (3) To further improve the correct rate of disease diagnosis, research-based the apple disease diagnosis of BP neural network model, the dynamically generated code, diseases, diagnostic parameters stored in the knowledge base medium strategy for network programming, parameters of disease diagnosis input, disease diagnosis results output, BP neural network in Web mode using Apple's disease diagnosis. Experimental results show that the three groups of test samples of BP neural network, the correct diagnosis rate reached 93.30%, 80.00%, 46.70%, were higher than the diagnostic accuracy of the similarity model. (4) the diagnostic algorithm evaluation system was designed and implemented, as well as a three-tier B / S mode apple disease intelligent diagnosis system. For the apple diseases intelligent diagnosis system architecture and functional modules were designed the algorithm testing software and disease diagnosis software, the use of the algorithm testing software testing, validation and demonstration algorithm to calculate the proposed diagnostic method, but also allows users to Knowledge Base knowledge to add, delete, and modify, apple disease intelligent diagnosis system completed queries disease, diagnosis, prevention and treatment guidance, expert exchange.
|