Research On The Application Of Agricultural Internet Of Things And Its Data Fusion Technology

The concept of the Internet of things was proposed by the Massachusetts institute of technology in 1999,then the Internet of things technology developed rapidly,it has been widely used in industry,medicine and safety engineering,agricultural Internet of things is one of the important development directions of Internet of things,The application of Internet of things technology in agriculture will bring great impetus to the development of agriculture and improve the intelligence level of agriculture.The data produced in the process of agricultural production is characterized by a large number,complex structure,various types,low value density,and rapid production.Therefore,it is necessary to analyze the data fusion in the production process.This paper completed the iot farmland information collection platform design and build,the use of ZigBee sensor technology and related agriculture construction of wireless sensor network,as a field microclimate data acquisition module,using China meteorological data sharing service platfonn provided by the API interface to get meteorological data as farmland climate data acquisition module,then use GPRS and 3 g technology to complete data transmission,the complete remote terminal PC web publishing.The temperature and humidity,C02 concentration,light intensity,soil moisture and soil temperature of the experiment field were monitored in real time through the installation of field IoT system in a Pitaya experimental field of Irrigation Experiment Station in Nanning.First,the data obtained from the experimental field were cleaned,and then the environmental data of each group were marked as "suitable" and "unsuitable" according to the requirements of the growth environment of pitaya.Finally,the data were automatically classified by the decision tree method,and the correct rate was used as an indicator to evaluate the model performance.The experimental results showed that the correct rate of the test set of the decision tree was 99.04%,and the correct rate of the verification set is 100%.It showed that the decision tree had a good performance in data classification.In the last part,the prediction model of farmland pests based on d-s evidence theory based on artificial neural network is proposed,Using average temperature,minimum temperature,rainfall,sunshine duration,4 common and easily accessible climatic data as input,The occurrence degree of insect pest is the output quantity,Respectively using BP neural network forecasting model and Elman neural network prediction model of training sample data and predicted,the predicted results and improvement of D-S evidence theory based on artificial neural network model for prediction of the prediction result is compared,the results show that based on D-S evidence theory of artificial neural network prediction model of prediction accuracy is much higher than before the artificial neural network prediction model.It can better reflect the occurrence of pests.