Building Of Warning System On Greenhouse Roses Diseases Based On Genetic BP Netural Network

Rose diseases has become an important question,which effect rose production and development,lead to low efficiency and quality of rose products,the lack of competitiveness in the international market.Therefore,how to forecast rose diseases accurately and effectively,become a top priority in the rose production management.In recent years,the artificial neural network were widely used in the plant diseases and insect pests,has received the good effect.In this paper,according to pathogenesis regularity of rose diseases,analyze the influence of environmental meteorological factors to diseases occurs,and build the BP neural network model,which determines six great influence factors of rose diseases:minimum temperature,maximum temperature,average temperature,minimum humidity,highest humidityand average humidity,this six indexes as network's input layer,three indicators as the output of the network,they are rose powdery mildew,downy mildew,gray mold.But because of the BP neural network easily fall in local optimumand cann't convergence or slow,so put forward to optimize and improve the BP neurral network of using genetic algorithm,give full play to the advantages of both the BP neural network and genetic algorithm,bulid the prediction model of rose diseases based on genetic algorithm and BP neural network.In order to compare two model,use Matlab2009b to analysis and contrast model result.The results show that the model based on genetic algorithm and BP neural network can predict the diseases of rose,and compare to BP neural network,t-he model based on genetic algorithm and BP neural network is superior to traditional BP network model,which have high precision and stability.Normal physiological function of contaminated with rose powdery mildew will be certain degree impacted, through the CI-310 determination system of portable photosynthesis,we measure net photosynthetic rate of rose leaves for health leaves and contaminated with rose powdery mildew, using plants effectiveness analyzer (PEA) measure fluorescence parameters Fv/Fm rose leaves for health leaves and contaminated with rose powdery mildew. The results show that:along with the increase of disease strength, net photosynthetic rate and fluorescence parameters Fv/Fm of rose leaves reflect a downward trend,and establish the regression equation among net photosynthetic rate,fluorescence parameters Fv/Fm of rose leaves with sense index of powdery mildew disease,the decisive coefficients R净2 =0.95, R荧2 =0.972.it show that we can use The photosynthetic characteristics instead of traditional monitoring method for feeling disease index of rose powdery mildew.Based on the research of rose diseases model,using of hybrid programming of Matlab2009b and Visual C ++,build warning system of rose diseases,so users quickly inquires the related knowledges of rose powdery mildew,downy mildew,gray mold,such as pathogenesis regularity,symptoms and prevention measures,at the sametime,according to prediction system can forecast possible diseases of rose,so users enable to early detection and early prevention.The result showed that the system can evaluate and better predict diseases,which play a very active role to prevent rose diseases.