基于改进小波域阈值法的平移不变振动信号去噪

Translation-invariant denoising of vibration signal based on improved wavelet domain thresholding method

针对含噪振动信号的去噪问题,采用了目前最有效的小波算法。在传统小波域阈值法的基础上,克服了软、硬阈值的缺陷,采用了新的阈值函数,并通过平移不变小波变换对去噪效果进行了强化。通过与几种方法去噪效果的仿真对比,其结果表明,新的去噪方案可以获得最大的信噪比(SNR),其去噪效果明显优于传统的软、硬阈值函数,并在实际振动信号的处理中得到了很好的应用。

In order to denoise the vibration signal including noise, the most effective method of wavelet algorithm was adopted. The flaw of hard＇and-soft-threshold function was overcome on base of wavelet domain filtering method. The denoising effect in new threshold function was strengthened by cycle-spinning. The experiment results show that the new denoising scheme can carry out the best result in signal to noise ratio（SNR） valuation, compared with other known methods. And the scheme also can be used in practice.