Eggshell crack detection based on the transient impact analysis and cross-correlation method

To explore the transient impact process for cracked eggshell detection, an equivalent mechanical model was built based on a self-designed automatic excitation device. Through analysis of power spectrum from dynamic force signal, it was found that the impact speed affects only the impaction energy. In contrast, the material of the impact head and the weight of the excitation rod determine the energy and the cut-off frequency of the impaction. When the weight of impact tup is less than 6.62 g, the cut-off frequency of excitation impact can cover the egg's inherent frequency. Then, an optimized experiment system was designed to acquire the response acoustic signals. The cross-correlation analysis and Bayes classification methods were carried out to detect the cracked eggshell. In the conducted experiments, a crack detection level of 97% and a false rejection level of 1% were achieved. From the findings, it can be concluded that the proposed method will assist in optimizing the impact device and simplifying the classification algorithm for an online detection system.