数字超声系统脉动噪声的估计与消除

采用高通滤波器和小波包变换技术对低碳钢摩擦焊接头进行超声无损检测，研究表明，高通滤波后，信噪比有所提高，但频率滤波器在滤除噪声的同时，也滤除了缺陷信息中所含的低频成分。小波包变换是一种比小波变换频率线性度更好的时频分析方法，能同时兼顾信号的陡变和缓变特征，从而清楚地区分时变信号和长时间的类周期信号。实验证明，小波包换能正地估计出脉动噪声，从而有效地去除之，提高信号的可观察性和信噪比。

RINGING NOISE ESTIMATION AND ELIMINATION FOR A DIGITAL ULTRASONIC SYSTEM

High-pass filter and wavelet packet transform were utilized to estimate and eliminate ringing noise for ultrasonic testing of the friction-welded joints in low carbon steel. It was indicated that high-pass filtering of signals led to enhancement of signal-to-noise ratio(SNR) to some extent, but at the same time, when ringing noise was filtered by frequency filter, the low frequency component of defect echo was filtered out, too. Being a time-frequency analysis with a better frequency linearity than wavelet transform, wavelet packet transform took care of both choppy and trend characterization of a signal, and distinctly identified different components of an ultrasonic signal which was a superimposition of short time-variable component and quasi-periodical long-term component. Experiment showed that wavelet packet transform seemed perfect for estimating ringing noise, and the elimination of the noise led to the enhancement of SNR and flaw visibility.