基于小波收缩策略的一种新阈值去噪方法研究

Research of New Threshold Denoising Method Based on Wavelet Shrinkage

分析了信号处理中相干噪声的特点,针对小波变换标准软阈值去噪法处理相干噪声能力相对差的弱点,介绍了一种基于SURERisk理论建立的自适应阈值去噪法,但其数字计算能力较差。基于同样理论,提出一种新的阈值函数,具有二阶导数和实际应用数字特性。并在此基础上确定了自适应阈值计算方法,建立了新的阈值去噪法,可用于去除相干噪声。计算机仿真表明,对于受相干噪声干扰的数字信号,该方法能较大提高信噪比,较好恢复信号。

In this paper, we analyze the coherent noise contaminating the signals, which can be weakly restored by the standard soft threshold denoising method. For this, we introduce a proved adaptive threshold denoising method derived form SURE Risk, but it has a little bit poor numerical property. Based on the same theory, we present a new thresholding function with continuous second order derivatives and practical numerical property. And then, establish the adaptive thresholding algorithm. Thus, we get a new adaptive threshold denoising method fit for coherent noise. The computer simulation is made, and the results show that this method get higher signal to noise ratio and better signal restoration, when apply to the coherent noise corrupted digital signal.