Development Of Web-Based Expert System For Comprehensive Control On Pear Valsa Canker

Pear,as the third largest fruit in our country after apple and orange,has quickly developed in recent years.China contained abundant germplasm resource,and the cultivated area and total yield all among the world's largest.Now pear industry is the important pillar that promotes the agricultural development,rural economic prosperity and peasant rich in main producing area.Valsa canker is one of the most serious trunk diseases in pear planting and wide spread occurred.The affected plants may subsequently rot bark,decrease both yield and quality,and in severe cases can cause death or orchard-destroy.Because of its high incidence,extensive distribution and hard to be controlled,pear canker is fairy difficult to be prevented and definitely hindered the healthy development of the pear industry.However,the current prevention and prediction mainly depends on famers' personal experience.What's more,existing experts from related fields far from enough to meet the growing the demand of famers and managers.This article tries to develop a web-based expert system for comprehensive control on pear canker,and build the prediction model of pear canker based on BP neural network.The system has three main functional modules: the recognition of pear canker,the epidemic law of pear canker and related comprehensive control.The mainly research in this paper is the following:1.To confirm the influence factor of BP neural network predictive model.With the incidence of pear canker in Yangling regions as basis,year-round releasing of conidia,year-round spreading out of disease spot and the year-round fluctuation of pear canker are analyzed and discussed.Combined with the local climate,we also studied the epidemic law of pear canker and correlation between meteorological factors and the development of pear canker to determine proper input variable for BP neural network.2.To build a BP prediction model of pear canker.At first,principal component analysis(PCA)was used to prepare and analyze the data in advance.Then use MATLAB neural network toolbox to train the experimental data and perform simulation test to compare results between the prediction and experimental data.3.To develop an expert system for comprehensive control on pear tree canker by using computer technology and network information techniques.The B/S-based system which provides users with a convenient and ready-to-use web services platform was programmed in html,CSS and JavaScript by Sublime and Dreamweaver.The advantages of diversified network techniques and the architecture of B/S gave this system extensibility,friendliness and dynamics,which not only enriches the content of system and human-computer interface,but also make it easy to maintain and real-time update.By analyzing these questions,temperature,rainfall and humidity are key factors to affect the occurring and spreading of pear canker.Its predictive model based on BP neural network may improve the accuracy by improved algorithm.Compared with other mathematical models,BP neural network can solve nonlinear problems.The development of web-based expert system for pear canker provides users scientific control measures and remote assistance.It not only has characteristics such as information sharing and practicability,but also provides reference for further study.