The Prediction And Research About Grain Stored Insect Base On The Chaos

The chaotic systematic analysis is applied to the field of grain pests’ prediction research in the article firstly. Its characteristics is used to analyze the impact factors of grain pests start with the article. Then analyzing chaotic characteristics of the stored grain insect pests based on experimental data, deeply discussing the combination of chaos and BP neural networks and support vector machines, establishing forecasting model of pests. In a word,the main research contents and results are as follows:1. Analyzing the main factors of grain pests. That four main influencing factors of temperature, humidity, moisture, gas and their mathematical models are analyzed detailedly.Then utilizing interpolation method to simulate the cloud picture of grain temperature and humidity, which efficiently monitors the trend of temperature and humidity variation, laying the foundation for further judgment pests;2. Analyzing Sitophilus zamia’s chaotic characteristics during the process of grain pests’ occurring a comprehensively. By analysing these features of Autocorrelation Function,Partial Autocorrelation Function, Power Spectral Function, Lyapunov index, Principal Component Analysis and the Poincare, the results can be determined that the time occurring series of Sitophilus zamia is chaotic state,which is conform to specific rule and is not completely random;3. Studying the BP neural network prediction model of chaotic sequences and SVM prediction model of chaotic pests’ detection and classification. In the BP neural network,combined with the evaluation of prediction model and stored grain pests situation index,basic research is analyzed about the model analysis, construction and forecasting of chaotic time series, the results show that BP network prediction of stored grain pests(such as Sitophilus zamia) is feasible.In the pests’ classification model research based on support vector machine, designed the collection device to gathers and analyzes the sound signal of Rhizopertha Dominica and Sitophilus oryzae Linne. Then the article proposed storage pests’ detection and classification method based on chaos optimization algorithm and SVM, which a series of simulations was accomplished to verify the feasibility of identifying different types of pestssound signal.To sum up, in this paper, new theories and methods were applied to study grain pests’ prediction, which is a positive and effective exploration to improve forecasting accuracy rate about grain pests.