Non-destructive assessment of hens' eggs quality using image analysis and machine learning

Abstract

Eggs are an essential source of inexpensive protein. Due to their oval shape, eggs can be efficiently handled, transported, and packed. The geometric description of the egg shape has been used as another parameter to evaluate the quality. Egg quality traits are related to several variables, such as the rearing environment, nutrition, breed, and age of the hen. We hypothesize that the shape index is associated with egg quality traits and that its isolated analysis can be used in the egg classification process. Given the complexity of variables affecting egg quality traits, we believe that knowing how the internal and eggshell quality relates to its shape may favor the classification process. Our study analyzed the associations between egg shape (using Shape Index, SI) and quality traits. We tested several machine-learning models to establish a relationship between shape and egg quality traits. From the images of 6,378 eggs, we found rounder eggs (SI ≥ 76) to have internal and eggshell quality higher than more elongated eggs (SI < 72). The best fit model was the Random Forest, with an accuracy of 97.9%. Assessing egg quality using a non-destructive method based on image analysis of egg shape can improve the grading process of commercial eggs in the processing industry.

Keywords

• Precision food processing
• Mathematical modeling
• Non-destructive measurements
• Laying hen