Mildew Prediction of Warehousing Tobacco Based on BP Neural Network
There is still no effective solution to mildew prediction. In order to forecast the moldy degree of ware-housing tobacco in real time,a tobacco mildew prediction model is established by BP neural network,in which temperature,humidity,and tobacco moisture are selected as the network input and the mildew degrees of the to-bacco are extracted as the network output. Firstly,the measured data of 78 sets are used as training samples to obtain the threshold value and the weight value of BP neural network. The data of 14 samples are simulated with linear regression analysis to validate the proposed model. The results show that the deviation range be-tween the predictive value and the actual value is[-0.028,0.033],and the relative error’s absolute mean is 0.0019 . Finally,the tobacco mildew real-time prediction is proved to have higher prediction precision in the to-bacco warehouse intelligent monitoring system based on embedded ARM+Linux+Web.
国家科技期刊平台
NSTL
万方数据
维普
