Prediction Of The Pine Wood Nematode Based On Artificial Neural Network And Hyperspectral Data

Pine wilt disease caused by the pine wood nematode is a highly hazardous international quarantine diseases.The disease has been found in the Mausoleum of Nanjing since 1982,which spread rapidly and prevent difficultly in the country.Hyperspectral imaging technology obsessed a continuous band, narrow spectral range, data volume and other advantages, featuring high resolution, while the advantages of artificial neural networks in the classification prediction has also been widely recognized. Hyperspectral data classification using artificial neural networks as forecasting methods in agriculture and forestry production has a certain application. In this study, Pinghu black pine and pine wilt disease in Yuyao massoniana affect the normal and various pathological hyperspectral data, after data analysis by variance analysis method to analyze hyperspectral data, choosing in favor of the experiment carried out data, using a variety of artificial neural network model to classify forecast.The result showed that there are higher average accuracy during prediction of BP neural network, RBF neural network, and Elman neural network, we can forecast correct and various pathological pine wilt disease, and integrated forecasting accuracy time, stability and other factors to described that above three neural networks have their own advantages in the pine wilt disease prediction.