| Barley yellow dwarf virus has occurred the most common and catastrophic crop disease in Shaanxi in relate years. It affects gramineous hosts by causing severe disease on wheat during seedling stage. Apparently, it is more important to enforce forecasting of impending epidemics in the current circumstance that is lack of effective resistant varieties and chemical curing methods. In the same condition, apple marssonina leaf spot has become more serious disease which caused by apple marssonina blotch because of apple bags has widely used and not enough attention has been prayed on the before-harvest apple diseases . Apple production which is concerned as one of the six pillar industries of Shaanxi province. Marssonina mali , often occurs in large area and reduces over 85% orchard production when epidemic year .When Apple marssonina blotch happens, it occurs abundant leaf drops and keep its negative influence for several years. Forecasting of apple marssonina blotch has merely descriptive reported which is lack of model supported. The forecasting models which been used at moment are using linear model prediction, this kind of model only can analyses linear factors of influence before the disease occurs. However, most diseases processes are very complex, time, environment, and the relationship between meteorological factors that should all be concerned which is not a simple linear relationship. Therefore we choose two main cultivated crops in Shaanxi province, use two different kinds of nonlinear mathematical modeling methods to analyze the nonlinear model in the prediction of rationality in disease, in order to extend the way to solve the agricultural diseases forecasting issue. According to the objectives above, we establish the Shaanxi wheat dwarf BP artificial neural network model with nonlinear apple marssonina blotch polynomial fitting model. Main contents include: Wheat dwarf BP neural network model methods and establishing, Apple marssonina blotch harm degree of Shaanxi province of Shaanxi province, influence, the popular apple marssonina blotch environment and meteorological factors, Apple marssonina blotch dynamic prediction model construction and application. Main research is as follows:1. Using the laboratory monitoring data of wheat dwarf which has been collected in Shaanxi province in the late 20 years (1991-2009), it demonstrate the disease happened every year in different degree. We submit the modeling analysis should be related to 23 impact factors about the wheat dwarf.2. We set up a complete construction of BP neural network methods, and establish the wheat dwarf BP neural network model. The rationality of the model of nonlinear analysis on factors and high sensitivity, illustrates the nonlinear model in the prediction of disease, also develop the BP neural network model of application.3. Apples to apples, cultivation of Shaanxi marssonina blotch serious harm, investigation and study the apple of Shaanxi province marssonina blotch harm degree, and environmental factors of nearly ten years apple marssonina blotch, prevalent regularity and in-depth analysis, temperature, humidity, wind speed, varieties and human management situation of apple, marssonina blotch influence extent of popular.4. Matlab software polynomial fitting construct apple marssonina blotch dynamic prediction model, analyses the model that the forecast result and practical survey data has good compliance. Through the model test apple marssonina blotch new disease spot conditions for temperature, humidity and temperature, big occurs 50% of the 23, humidity levels. Put forward a new method of forecasting model is established, and also widened the mathematical modeling in forecasting methods of plant disease areas.
|
PDF Full Text Request