Automatic classification of magnetic tiles internal defects based on acoustic resonance analysis

• A new method for automatic detection of internal defects in magnetic tile.
• The acoustic signal is obtained by mechanical system based on acoustic resonance.
• WPT, LDA and SVM are combined to analyze the acoustic signal.
• The constraint algorithm can effectively improve the accuracy of identification.
• The presence or absence of internal defects can be identified with great accuracy.

A novel signal processing method using wavelet packet transform (WPT), linear discriminant analysis (LDA) and support vector machine (SVM) is presented for detecting internal defects in magnetic tile. In this methodology, the acoustic signal obtained by a mechanical system based on acoustic resonance is analyzed. WPT is applied to extract the normalized features of the signal. The internal defects are identified by SVM based on the extracted features optimized by LDA and a constraint algorithm. The experimental results demonstrate that the presented approach can be employed for a promising application of automatic detection of internal defects in magnetic tile.