近红外光谱技术结合反向区间偏最小二乘法预测水稻植株氮含量OA

To realize real-time monitoring of nitrogen(N)content and fertilization regulation during the process of rice growth,a method for rapidly determining N content in rice plants based on near infrared spectroscopy was proposed.A near infrared spectrometer was used to collect the near infrared spectra of samples,and N content was determined by the Kjeldahl method.Raw spectral data were pretreated by standard normal variate(SNV),first derivative(FD),and their combined method.Backward interval partial least squares(BiPLS)was employed to select characteristic intervals,thus achieving data dimensionality reduction and model simplification,and a partial least squares(PLS)prediction model was established.The results showed that the model established with 402 characteristic wavelengths screened by SNV+FD pretreatment combined with BiPLS performed the best.The root mean square error of cross validation(RMSECV)and correlation coefficient(Rc2)were 0.197 9 and 0.952 5,respectively.The root mean square error of prediction(RMSEP)and correlation coefficient(Rp2)were 0.201 3 and 0.950 6,respectively,and the residual predictive deviation(RPD)was 3.22.The study demonstrated that the near infrared quantitative analysis model established by the SNV+FD pretreatment and BiPLS exhibited excellent accuracy and stability.It is suitable for the rapid monitoring of N content in rice plants and provides a technical reference for the prevention of rice diseases and pests and scientific field management.