小波变换信号消噪技术研究

Research on the Signal Denoising Technology with Wavelet Transform

通过对信号与噪声奇异性分析,给出信号与噪声的小波变换模极大值在各个尺度上的表现截然相反的结论,并给出了一种非线性的消噪方法.该方法与传统的消噪方法不同,即不同于低通滤波器滤波,而是根据信号与噪声的奇异点性质不同进行滤波.实验结果表明,这种方法在改善信噪比的同时,又保持相当高的时间分辨率,并且计算简单,有很好的去噪效果.

Wavelet transform is well localized characteristic both in time and frequency field. The singularities of signals and noises are investigated. A conclusion is drawn that the wavelet modulus maxima of signals have contrary behaviors from that of noises. At this foundmation, a nonlinear noise eliminating method is proposed. Distinguished from the traditional method, it is not equivalent to letting the signal pass a low-pass filter, but carrying out the filtering on the basis of the different properties of the singularity of the signal and noise. The experiment shows that the method proposed not only improves the signal to noise ratio but also makes the time resolution remain high, and can suppress the noise effectively.