A wavelet analysis-based matching pursuit algorithm for an accurate ultrasonic TOFD measurement

The time-of-flight diffraction (TOFD) technique is used as an important non-destructive testing method in weld integrity evaluation and failure analysis. However, an accurate measurement of the time-of-flight (TOF) has proven to be difficult due to the low time resolution of the measured signal. Conventional deconvolution techniques have been used to improve the time resolution of the signal but are not effective for ultrasonic TOFD signals because the frequency contents of the signals are non-static in space-frequency distribution. To overcome this problem, a method is proposed in this paper that estimates the TOF in two steps. In the first step, the measured signal is decomposed into a series of narrowband signals using a wavelet transform and an atom dictionary is adaptively established according to the characteristics of a selected narrowband signal. In the second step, matching pursuit (MP) is used to derive a sparse representation of the selected narrowband signal. A steel specimen with artificial defects is prepared, experiments are carried out and the results confirm the efficacy of the proposed algorithm.