Higher-order statistics-based deconvolution of ultrasonic nondestructive testing signals

Pulse-echo reflection techniques are used for ultrasonic flaw detection in most commercial instruments. As the measured pulse echo signal is assumed to be the result of linearly convolving the defect impulse response (IR) with the measurement system response, the objective is thus, to remove the effect of the measurement system through a deconvolution operation and extract the defect impulse response. The major drawbacks of conventional second-order statistics (SOS)-based deconvolution techniques are their inability to identify non-minimum phase systems, and their sensitivity to additive Gaussian noise. Our contribution is to show that higher-order statistics (HOS)-based deconvolution techniques are more suitable to unravel the effects of the measurement systems and the additive Gaussian noise. Synthetic as well as real ultrasonic signals are used to support this claim.