We previously showed in part 1 that a near-infrared spectroscopic sensing system constructed on an experimental basis can be used to assess milk quality in the laboratory. In this study, we developed calibration models for predicting three major milk constituents (fat, protein and lactose) during milking by the sensing system and validated the precision and accuracy of the models. Coefficient of determination (r
2) and standard error of prediction (SEP) of the validation set for fat were 0.94 and 0.55%, respectively. The values of r
2 and SEP for protein were 0.80 and 0.13%, respectively, and the values of r
2 and SEP for lactose were 0.83 and 0.09%, respectively. These results indicate that the sensing system can be used to assess milk quality in real-time during milking. From the data of milk constituents measured by the sensing system and milk flow rate during milking, we calculated milk constituents of whole one milking. The calculated values of milk constituents well agreed with the reference values.
View full abstract