NIR Spectroscopy Applied In Rapid Analysis Of Moisture And Fat In Corn

Corn is very important, as a human food and raw material feed. After companies purchase this material, the protein, moisture and fat contents usually were tested by the conventional detection methods. Now, the conventional detection methods have so many disadvantages that need preparing reagents, complex operations, and long time. In this paper, near infrared diffuse reflectance spectroscopy and chemo metrics methods which can be applied in analyzing moisture and fat of corn flour quantitatively and rapidly without reagent.Collected samples would be randomly divided into calibration, prediction, validation sample set. Calibration sample set was used to determine the parameters of the near-infrared quantitative models; prediction sample set was used for quantitative assessment model; the preferred model parameters were determined according to forecast choices; Finally, testing the preferred model by the validation sample which was not involved in the modeling process set.Equidistant combination partial least squares(EC-PLS) can be used to establish near-infrared wavenumbers models about high signal-to-noise ratio, and the models which could be used to test moisture and fat of corn flour. For the moisture analysis, Beginning wavenumber(B), Number of wavenumbers(N), Gap of wavenumber(G) of the optimal EC-PLS model is 5785 cm-1, 31, 8, respectively; for the validation sample, the root mean square error prediction about predicted values by near-infrared and measured values by conventional detection methods(V-SEP), the correlation coefficient of prediction(V-RP) for moisture were 0.278%, 0.943, respectively. For the fat analysis, the best B, N, G is 4568 cm-1, 22, 10, respectively, validation samples’ V-SEP, V-RP were 0.253%, 0.891. The results showed that the near-infrared predicted values are relevant to the measured values by conventional methods, which meet the precision requirements of feed ingredients near infrared analysis from GB/ T18868-2002.The method was a potential and new technology which was suitable for large-scale testing of feed ingredients because of the advantage of no reagents, quick and easy. In this paper, wavenumber options about anglicizing water and fat of corn flour can provide valuable reference for the design of small special near-infrared instrument.