NIR Spectroscopy Combined With EC-PLS To Establish The Joint-Analyses Model For Multiple Quality Indicators Of Corn

Near infrared(NIR) spectroscopy has been successfully applied in many areas, with several advantages in real-time, quick and nondestructive detection. Based on the sample information, spectroscopy quantitative analysis is the method of determination of composition of the samples with computer mathematics modeling method. The commonly used modeling analysis methods are: principal component analysis(PCA) and partial least squares(PLS), moving window partial least squares(MW-PLS), artificial neural network(ANNs) and support vector regression(SVR) analysis method and so on. Modeling, to obtain the characteristics of the high signal-to-noise ratio wavelength is both a key, is also the difficulty. Based on the EC-PLS, mathematical model for multi-index joint-analyses of NIR quantitative analysis is proposed in this paper. And the method was successfully applied to multi-index joint quantitative analysis of crude protein, moisture and crude fat in corn.One hundred fifty-six corn samples were collected, a random and representative framework of calibration, prediction, and validation was proposed. All the samples were randomly divided into model set and test set. Calibration set is used to determine the model parameters; Prediction set was used to evaluate model prediction effect; then, according to the prediction effect to determine the optimal model parameters; finally, the model is validated by the validation set which is not involved in modeling. The equidistant combination partial least squares(EC-PLS) method was improved basis on above, the NIR quantitative analysis of independent and joint-analyses models with high signal-to-noise to crude protein, moisture and crude fat in corn were establishment respectively, and comparison with their effects.Random validation samples excluded from the modeling process were used to validate the four selected models. For the independent analysis models, the validation root mean square errors(V_SEP), validation correlation coefficients(V_RP), and relative validation root mean square errors(V_RSEP) of prediction were 0.271%, 0.946, and 2.8% for crude protein, 0.275%, 0.936, and 2.6% for moisture, and 0.183%, 0.924, and 4.5% for crude fat, respectively. For the joint-analyses model, the V_SEP, V_RP, and V_RSEP were 0.302%, 0.934, and 3.2% for crude protein, 0.280%, 0.935, and 2.7% for moisture, and 0.228%, 0.910, and 5.6% for crude fat, respectively. The Results show that the near infrared predicted values and measured values has good congruency and high correlation, and infrared predicted possesses the advantages of no reagent, quick and easy, also suitable for test requirements in large-scale agricultural production.In this paper, the wavelength of the model can be successfully applied in multi-index quantitative analysis in crop or feed, and the proposed wavenumber selection method provided also valuable reference for designing small dedicated spectrometer for corn.