Maximum nongaussianity principle for ultrasonic flaw detection

Maximum nongaussianity principle for ultrasonic flaw detection Hua-ming LEI, Wei LIANG, Pei-wen QUE Department of Instrument Science and Engineering, Shanghai Jiaotong University, Shanghai, 200240, P.R.China E-mail: [email protected] Abstract: The noise suppression techniques with wavelet transform (WT) are widely used in non-destructive testing and evaluation (NDT&E). In this paper, we have analyzed the non-gaussian statistical properties of the ultrasonic echoes. The distribution of the sum of the M-superimposed Gaussian echoes and WGN is closer to gaussian distribution than that of the M-superimposed ultrasonic Gaussian echoes. The smaller the signal to noise ratio of the M-superimposed Gaussian echoes, the weaker the nongaussianity of the ultrasonic Gaussian echoes. In this paper, a technique for improving the signal to noise ratio of ultrasonic signals using optimal frequency-to-bandwidth ratio wavelet transform (OWT) is presented. We address the design of this wavelet using maximum nongaussianity principle. Experimental results have been presented to evaluate the effect of the optimal wavelet transform filtering on ultrasonic flaw detection. All results showed that the optimal wavelet has best effect of the noise suppression and in much improved detection of the signals.