Ultrasonic Flaw Echoes Detection Based on Generalized S-Transform

Abstract

Micro defects usually lead to performance degradation of materials. Pulse-echo technique is an effective method of revealing micro defects in materials, and booming in which has been witnessed over the years. However, detection of flaw echoes submerged in intensive noise is challenging. In this paper, a method of flaw echoes detection based on generalized S-transform is proposed. The original signals are transformed to time-frequency domain by generalized S-transform in the first stage. Subsequently, the time-frequency spectrums are enhanced through image segmentation techniques, including global thresholding and dilation. Finally, clear flaw echoes are obtained by inverse generalized S-transform, contributing to further investigation of defects. Both the simulation and experimental results show effectiveness of the flaw echoes detection method.