Ultrasonic nondestructive evaluation of highly scattering materials using adaptive filtering and detection

Adaptive filtering and detection has been applied to the problem of detecting ultrasonic echo signals from test targets where the wanted signals are masked by coherent scattering from grain boundaries present in highly scattering materials. The filter is based on the normalized least mean square (LMS) error algorithm, and can be operated with either an independent reference signal or by using the delayed input signal as the reference. Tests made on a collection of 64 ultrasonic A-scans using the same processing parameters show that an up to 10 dB improvement in signal-to-noise ratio can typically be obtained. A cell-averaging constant false alarm rate (CFAR) detector is used to detect the signals automatically. The performance of the method is compared to that of split spectrum processing, both with and without polarity thresholding.< >