Wavelet based noise suppression technique and its application to ultrasonic flaw detection

Area of Science:

• Non-destructive Testing and Evaluation (NDT&E)
• Signal Processing
• Ultrasonic Testing

Background:

• Wavelet transform (WT) is common for ultrasonic signal noise suppression in NDT&E.
• Existing WT methods include thresholding, singularity analysis, and matched filtering.
• A simpler, more effective noise suppression technique is needed.

Purpose of the Study:

• To develop a new, simple, and effective WT-based noise suppression technique for ultrasonic signals.
• To treat WT as a tunable band-pass filter for optimal noise removal.
• To evaluate the technique's performance in challenging, noisy conditions.

Main Methods:

• WT is employed as a band-pass filter, with central frequency and bandwidth (CF&FB) derived from signal spectra.
• An optimal scale for the daughter wavelet is determined to match the signal's CF&FB.
• The denoised signals are analyzed using computer-generated white noise and experimental pipeline flaw detection.

Main Results:

• The technique achieves excellent denoising results, even with signal-to-noise ratios (SNR) below -18 dB.
• Performance was validated using ultrasonic signals with simulated white noise.
• Experimental verification on a pipeline specimen with artificial flaws yielded positive outcomes.

Conclusions:

• The developed WT technique offers a simple and effective solution for ultrasonic signal noise suppression.
• It is particularly suitable for processing heavily noised signals in NDT&E.
• The method shows potential for application in automatic flaw detection systems.