The Construction And Implementation Of Tomato Botrytis Cinerea Early Warning Model Based On Rough Neural Network

The facility agriculture is ever-growing as a share of a total agricultural area and it is an effective project to solve the problem that more people and less land restricts sustainable development. However, the development of facility agriculture is faced with a difficult problem that the increasingly serious situation of facility agriculture diseases and insect. At present, the monitoring data collection method of the disease information mainly relies on field investigation carried out by the plant protection staffs, sampling and analysis, so the data size is not enough and directly affects the accuracy, timeliness and effectiveness of the early warning analysis. According to various environmental parameters of the greenhouse tomatoes which are real-time gathered through the Internet of things technology, the tomato Botrytis cinerea early warning model based on rough neural network is built and it can analyze the real time gathered data and then forecast tomato Botrytis cinerea. The key research contents of the paper are as follows:1. In the greenhouse environment, variety of real time data is gathered through various kinds of sensors in perception layer of the Internet of things and it mainly includes the temperature and humidity of the air and soil, light and carbon dioxide concentration, which is stored in temporary databases of the cloud platform and Hadoop cluster.2. The early warning model construction of tomato Botrytis cinerea. The paper firstly studies the effect that the double environmental factors impact on tomato Botrytis cinerea and gets the result that there exists certain limitations and then makes further research based on multi-factors. Because the number of the factor is more, we combine respectively traditional BP neural network and improved BP neural network and rough set theory. Through the instance analysis, the results show that the precision and training time of the early warning model based on rough neural network which is provided in the paper both have certain enhancement. The establishment idea of the model can provide beneficial references for the research of the early warning of disease and insect.3. The research and development of the early warning systems of tomato Botrytis cinerea. Based on early warning model which has been built, JAVA language, MyEclipse development environment and SpringMVC architecture, the early warning system is developed. The real-time data is stored in MySql and Hadoop cluster, JS and JQuery and other related technologies are used for front page display.Based on the research the paper takes tomato Botrytis cinerea as example and develops tomato Botrytis cinerea early warning system based on rough neural network. The system is part of the infrastructure cloud services in the SAAS layer of the agricultural cloud service platform of agricultural city which is constructed by national modern agricultural city for science and technology. And it is deployed and runs in the agricultural cloud service platform of agricultural city and its running stability is well. The above research not only can be used as reference materials to build other early warning systems, but provides offer practical significance for the realization of the green plant protection of the facility agriculture.