[Egg quality prediction by using Fourier transform near infrared reflectance spectroscopy (FT-NIR)]

The objective of the present study is to investigate the feasibility of using FT-NIR to determine egg quality: egg albumen height (EAH); Haugh unit (HU); air cell height (ACH); air cell diameter (ACD); egg shape index (ESI) and egg weight (EW). All eggs were stored in the same environment with 16 degrees C and 75% relative humidity after collection. Calibrations were developed using a leave-one-out cross validation procedure under partial least-squares regression. The different optimal parameter comparisons showed that Savitzky-Golay filter method and second derivatives can give the best calibration for all the measured values. The partial least-square analysis showed that R2 of the EAH, ACD and ACH are 0.867, 0.821 and 0.865, respectively. The RMSECV values are 0.476, 0.014 and 0.479 for EAH, ACD and ACH, respectively. The correlation coefficients of external-validation for 33 eggs are 0.873, 0.861 and 0.895 for EAH, ACD and ACH, respectively. The external-validation results are not significant different from actual measurements (P > 0.05) for EAH, ACD and ACH. All models cannot give the reasonable results for EW, ESI and HU. As the HU cannot be measured correctly by using FT-NIR and also is highly correlated with EAH, the authors suggested the possibility to revise the currently used measurement standard. Our results showed that FT-NIR could be used to test the egg albumen height, air cell height and air cell diameter.