Prediction for Available Nitrogen and Available Phosphorus by Using Hyperspectral Data

Based on the analysis of the laboratory reflectance spectra of 118 soil samples, the prediction model of the content of available nitrogen (AN) and available phosphorus (AP)was established by the method of partial least square (PLS). The results showed that after continuum removal and re-sampling, the correlation of soil nutrients (AN and AP) and spectral characteristics of soil had been improved. The determination coefficient R2 of calibration model of available nitrogen and available phosphorus established on this basis were 0.712 and 0.499, respectively. The determinative coefficients of validation model were 0.712 and 0.502, respectively.