Hyper-spectral inversion of soil Cu content based on BP neural network model
Based on the hyper-spectral data,for the reason of low fitting degree and poor prediction effect of the traditional inversion models of heavy models in soil,this paper extract the feature bands data of pretreated spectral for the correlation analysis,choose the first order differential spectral reflectance of 860 nm to establishes the BP neural network model of heavy metal Cu which based on Matlab,the fitting goodness of the model is 0.721 and the prediction accuracy is up to 82.3％,which were higher than those of the traditional unit linear regression model's fitting goodness of 0.414 and prediction accuracy of 76.1％.the study show that BP neural network model has better goodness of fit and prediction ability to predict the content of heavy metal Cu in soil more effectively.
hyper-spectralheavy metalBP neural networklinear regressiongoodness of fit
NETL
NSTL
万方数据
维普
