An improved denoising method for ultrasonic echo of non-destructive evaluation

As a widely used non-destructive evaluation (NDE), ultrasonic testing utilizes the echo wave to evaluate the material damage and degradation. However, there exists a lot of unwanted noise, which makes the detection of small faults difficult. In order to achieve the good detection performance of non-destructive evaluation, an improved denoising method of ultrasonic echo was proposed in present study. Firstly, the echo wave of ultrasonic testing is decomposed with the wavelet packet transform (WPT). Subsequently, the energy and entropy parameters of wavelet packets are calculated, respectively. Based on the analysis of their statistical properties, the threshold for the signal reconstruction packets is determined. Besides, by the simulation experiments on two cracks at different locations in a metal plate, the denoising performance of the present method using two popular wavelet functions of sym4 and db4 are investigated and compared with other conventional methods. All these experiment results show that the better denoise performance can be achieved with this proposed method especially when the signal noise ratio (SNR) is smaller than 10dB.