It is very important to detect the freshness of eggs. In order to achieve non-destructive detection of freshness, the electronic-nose technique was used in this study to predict the degrees of freshness of the eggs stored at a temperature of 20°C and a relative humidity (RH) of 70% by detecting volatiles. Meanwhile, as known indicators of the degree of freshness, the physical and chemical indices of eggs (Haugh unit and yolk index) were measured. The classification analysis of eggs stored for different numbers of days was conducted using linear discrimination analysis, and the result showed that these eggs could be distinguished effectively; the total contribution of the discriminant functions was 75.70%. Models of the relationship between the electronic-nose signal and the physical and chemical indices of eggs were established using multiple linear regression analysis and a back propagation (BP) neural network. The correlation coefficient and relative error of the multiple linear regression model were greater than 0.84 and around 8.00%, respectively. The correlation coefficient and the relative error of the BP neural network model were greater than 0.84 and around 9.00%, respectively. The results indicated that the electronic nose technique had a certain predictive capability of egg freshness, and this study can provide a reference for the non-destructive detection of the freshness of eggs.
