Analysis and simulation of noise reduction using wavelet transform for detecting defects in materials

The reflected sounds from the ultrasonic systems, which are being used to detect defects, are often affected due to the presence of noise. Reflections resulted from heterogeneous structures are commonly known as reverse distribution distortions. There are several methods to eliminate this noise. Wavelet conversion is one of the powerful and effective tools for noise cancellation. Mother''s wavelet compatibility with real signals on the one hand and less complexity of computations on the other hand make the wavelet conversion as an effective tool to be used. In this study, we used the wavelet transform method in order to eliminate the possible noise in ultrasound transducers. We applied a wavelet transform to eliminate noise in a defect identification system and, therefore, adapted the optimal algorithm for this purpose. Based on the simulation results and the utilized criteria of merit, the best results are obtained when the Meyer discrete mother wave, hard threshold and threshold level based on standard deviation are used. Our findings showed that the SNRE method, presented in this paper, improves the results considerably from 1 to 25 dB compared to conventional ultrasound noise reduction techniques depending on the percentage of recoil signal error amplitudes.