中文摘要：

提出了一种基于相关增强与小波阈值相结合的超声信号去噪方法,将该方法用于AZ91镁合金的超声检测能够有效地去除噪声,提高检测精度。由于同类缺陷信号之间相关性较大,因此对同类缺陷的超声信号进行相关处理能够增强缺陷信号;结合小波分析的多分辨分析特点,从频率和时域对增强后的信号进行阈值处理,最终获得了良好的去噪效果。实验结果表明：该方法能够有效去除由粗晶散射引起的结构噪声,对微弱缺陷信号的检测是很有效的,且灵活便捷。

英文摘要：

A signal processing method based on the combination of cross-correlation strengthening algorithm and wavelet threshold value was put forward. This method was used during ultrasonic testing of magnesium alloy AZ91 to remove the noise signal and improve the testing accuracy. The signals from the same kind of defects has much correlation, so the defect signals can be strengthened by cross-correlation algorithm. Besides, the threshold method of wavelet analysis is adopted to strengthen signal after using cross-correlation algorithm and good de-noising results are obtained. The results indicate that the noise caused by coarse grain scattering can be eliminated by using this method and this method can effectively detect tiny defects.