Research And Implementation Of Apple Diseases Intelligent Diagnosis System

Apple cultivation is a pillar industry in Shaanxi Province. One of the major factors that influence the quality and yield of apple is the comprehensive treatment to the disease and the insect pest. How to utilize the modern information technology to diagnose the disease and the insect pest and to supply some expeditious means to prevent and cure them becomes an urgent problem. According to the request of diagnosing, preventing and treating them, this paper researches primarily the acquirement of the disease screening knowledge and the inference method based on the analysis of the main characteristics of the disease and the insect pest of apple , and constructs the platform of apple disease knowledge base and diagnosis and the platform of intelligent system by introducing the diagnostic method based on similarity and BP Nerve Network. The major conclusions of this study are as follows:(1) Detailedly and comprehensively analyzes the factors that cause, effect apple diseases, species, knowledge types of apple disease. Conforms the diagnostic parameters of them and representation of apple disease knowledge. Adopts the method of dynamic encoding to generate automatically the parametric coding and the rule coding of apple disease diagnosis and stores them into the Knowledge Base. Studies the storage structure of the knowledge base. Sets up a knowledge base of apple disease based on the relation data base.(2) Researches the disease diagnosis model based on similarity. Defines the notion of apple disease diagnosis similarity. Gives the diagnosis method and criterion to apple disease similarity. Determines the share which the disease diagnosis parameter takes in the similarity diagnosis model. Establishes the similarity model and integrated platform. The experiment shows that the diagnosis accuracy rate based on the similarity model is 91.70%,78.40%,33.30%.(3) In order to increase the rate, researches the apple disease diagnosis model based on BP nerve network. Adopts the dynamic encoding and the strategy of designing the on-line programs through storing the disease diagnostic parameters into the knowledge base. Enters the disease diagnostic parameters and outputs the disease diagnostic results. Under the Web method, realizes the apple disease diagnosis based on the BP nerve network system. The experiment shows that the average accurate rate of the BP nerve network diagnosis can reach to 93.30%,80.00%,46.70%, which is higher than the similarity model. (4) Designs and realizes a diagnostic algorithm evaluating system and a 3-lyer B/S model intelligence diagnosis system. Towards the framework and the functional module of the intelligence diagnosis system, projects respectively an algorithm testing program and a disease diagnosis program. Uses the algorithm testing program to check the suggested diagnosis method, verify and demonstrate the computational process of the algorithm. What's more the user can add, increase and delete the data in the knowledge base and complete the process of checking, diagnosing, preventing and consulting the expert on-line by using the apple disease diagnosis system.