The Study And Development Of Forecast And Early-Warning Software System For Potato' Late Blight

The goal of this thesis is to develop a suit of Forecasting and Early-Warning software system for potato's Late Blight that can forecast the occurrence time and the trend of potato's Late Blight so that it can reduce the loss caused by Late Blight to a maximal extent.The main methods applied in the thesis are: a trend of potato's Late Blight of BP neural network long-term prediction model and Grey-Disaster prediction model are constructed with history data, and short-term prediction is constructed with the agriculture experts' experience. The three main modules: short-term prediction, BP neural network long-term prediction and Gray-Disaster prediction constitute the Forecasting and Early-Warning Software System for Potato's Late Blight.The main research gain of the thesis lies in:(1) The obtainment of short-term prediction model , BP NN long-term prediction model, and Grey-Disaster model of potato's Late BlightFirstly, the software system construct a history database related to potato' Late Blight; secondly, the short-term prediction model is constructed with the help of the experts' experience ; thirdly, a trend of potato's Late Blight of BP neural network long-term prediction model is constructed with the history data that is very helpful for the prediction of late Blight; fourthly, Grey-Disaster Model is constructed with the history data too that makes that we can predict the time of the disaster of late Blight, which has deep sense for our prediction work.(2) The Forecasting and Early-Warning software system of potato's Late Blight is developedThe system makes use of the Numerical-Computering of Matlab ToolBox function by Matlab Engine technology, whose user end is developed by Visual C++, and then exchanges data with Visual C++ to achieve the goal of predict. Meanwhile, the system build up the history database on the base of the DBMS:SQL Server 2000, which can be accessed by the MFC ODBC technology. Just as the above are mentioned, Forecasting and Early-Warning software system is developed. In fact ,it has been proved that the system is feasible. And the prediction method that combines the short-term prediction , long-term prediction of NN with Grey-Disaster prediction also may be applied for other domains to a certain extent, such as mechanic fault diagnosis, flood disaster prediction ,etc.