Eggshell crack detection using PCA and SVM

In the past, the inspection of cracks on eggshell was typically conducted in the industry by floodlighting, however, it gives eye fatigue, makes misjudgment and is not easy to detect hairline crack. Recent research into the automation of the detection of eggshell cracks is focused both on optical and mechanical detection principles. In this study eggs were excited by a light mechanical impact on different locations of the eggshell and vibrational frequency response of the eggshell combined with pattern recognition was attempted to differentiate intact egg and cracked egg. The pattern recognition was conducted by Principal Component Analysis (PCA) and Support Vector Machine (SVM). The optimal number of principal components was obtained 7 according to maximum error for predictive model and high discrimination between intact and cracked eggs. The result was found that the vibrational impulse response method distinguish intact egg and cracked egg with the level of 100% and 87.5% accuracy, respectively.