| The safety of food is an important task. Pests are the most important cause of grain losses. How to predict its density and to know the pest’s position have become the significant safe problems. Study the pest population density, which can effectively judge the growth trend of insects. Making sure the safety of food, we can make protective measures in advance.This essay expounds the research status of pest detection technique, prediction technology firstly, and then decrypts the development process, principle, distribution rules and application status of trapping technology. The correlations between the trapping amount and Grain ecosystem are analyzed. Finally, it is proposed predictive model of population density based on deep belief networks. The main research contents and the results are as follows:(1) Collecting basic data. Combined the requirement of food industry standard LST1203-2002, we draw the data of temperature, humidity and the population density from trap site and temperature-humidity integration sensor. Then three types of initial data are analyzed for correlation and density forecast.(2) The correlation between the trapping amount and temperature, humidity are analyzed. Firstly, obtaining the data of temperature and humidity in the grain bulk by sensor, and then getting the type and quantity of pests by trapping technology, researching the correlation between the trapping amount and temperature, and then the essay analyzed the relationship between temperature, humidity and pests trapping quantity by linear regression method.(3) This essay Studied grain pests’ growth trend prediction model based on Deep belief network. The deep belief networks forecasting model with 7 hidden layers is established firstly, simulating under the MATLAB environment, the accuracy of density population prediction is high. Furthermore, it demonstrated the method is a scientific, very quick and efficient new method of predicting the growth trend of pests. |
PDF Full Text Request