金属基复合材料缺陷超声信号消噪研究

Study on Multi -Resolution De-noising of Ultrasonic Signal From MMCs Using Wavelet Package Transform

为了剔除金属基复合材料缺陷超声信号中的噪声,得到没有污染的缺陷信号,便于进行信号分析,文章中使用了一种改进的小波包消噪方法——平均能量阅值法,讨论了平均能量阈值法消噪的原理,并将小波包的多尺度时频分析及重构结合在一起,用于信号的消噪研究中,通过仿真实验并与小波消噪方法进行对比,证明了该方法消噪的有效性;结果表明基于平均能量阈值法消噪技术在保持信号奇异性的同时,能有效地去除金属基复合材料缺陷超声信号中的噪声。

To eliminate noise in echo signal from MMCs and acquire purity defect signal, a modified method called average energy threshold using wavelet package transform is put forward for signal denoising. The principle of the wavelet package transform to reduce the noise is introduced. This method is incorporated with multi resolution analysis and reconfiguration to eliminate noise in echo signal. The a vailability of the method was verified by emulation experimentation and comparison with denoising using wavelet transform. By analysis the meaningfull denoising method to reduce ultrasonic signal is achieved.