Eggshell crack detection by using electrical conductivity and intelligent classification

Packaging, distribution and hatchery of the cracked in eggshell are always big challenges in egg industries. Due to the high volume of the egg production in the world, sorting eggs with cracks in the shell is so important. In this study, cracks in sell eggs are detected by using their electrical conductivity. In this study, 203 Hy-line eggs, including 48 micro-cracks egg were used. Since there is no measurement technique for detecting eggshell crack, 100 crack eggs were selected, and have been examined by expertise. Among them, 48 eggs with micro-cracks were selected and tested. For this purpose, dielectric constant and loss factor of eggs were measured between 40 KHz to 20MHz. These coefficients are then used as input matrix of classifiers, which are artificial neural network and support vector machine. 50% of data allocated for testing included 75 intacts eggs and 26 micro-cracks eggs. Both of the classifiers, Neural network and support vector machine, were detected cracks in shell eggs with 100% accuracy and 1 Kappa statistic.