| Summary: | Egg geometrical measurement is important for the poultry industry, and its calculation is not easily possible due to the unusual shape of the egg. To solve this problem a research has been carried out using a digital image analysis (IA) system to render the precise measurements of several egg size parameters, including egg volume (V) and surface area (S) of laying hen. We tested the accuracy of the IA method in determining egg physical properties by comparing the V resulting from IA with that measured using water displacement. The correlation of determination (R2) between the data obtained from these two methods was 0.98. We also applied the data sets of egg samples obtained by the IA to test the accuracy of the previously published equations to predict S and V in the egg samples. The results have shown that the equations posted by Carter (1975), Paganelli et al. (1974), and Narushin (1997) provided reasonable accuracy (R2 > 0.839) in predicting the egg S based on the length (L) and maximum breadth (B). In addition, the equations proposed by Carter (1975), Ayupov (1976), and Narushin (1994, 1997, 2005) provided accurate predictions for egg V by using L and B as the inputs. Furthermore, multiple linear regression (MLR), polynomial regression (PR), and artificial neural networks (ANN) models were used to test whether we could find new simple equations to predict the egg volume and surface area based on the egg weight, L, and B. The results indicated that weight could not be a helpful input variable, while weight is the single input of most published equations. Our newly developed models are also accurate for predicting V and S of egg samples based on L and B.
|
|---|