Edge detection of internal defects based on the hidden singularity of gradient streamlines obtained by continuous wavelet transform

Abstract

A method based on the measurement of out-of-plane displacement by the phase-shifting digital speckle pattern interferometry is proposed to detect the edge of internal defects. The out-of-plane displacement is scanned by the scanning method of gradient streamlines first, which can provide paths through the defect edge with the best signal-to-noise ratio everywhere. Then, continuous wavelet transform (CWT) combined with the weighted structural intensity of the wavelet maxima is applied to detect the hidden singularities of the gradient streamlines that form the defect edge later. Moreover, the weighted structural intensity method is optimized to adapt to the detection of the hidden singularity by CWT. The results of theory, simulation, and experiments agree well, and the proposed method is efficient in internal defect detection.