一种改进的小波阈值去噪法

An Improved Wavelet Thresholding Denoising Algorithm

基于小波变换的阈值去噪法是去除数字信号中白噪声的有效算法,其中阈值函数的选择关系着重构信号的连续性和精度。但是硬阈值函数在阈值点具有不连续性,软阈值函数中估计的小波系数与信号的小波信号间存在恒定偏差的缺陷,这些缺点限制了阈值去噪法的进一步应用。将硬阈值和软阈值函数进行加权平均,选取适当的权值函数,可以构造出一种克服了软硬阈值缺点的新的阈值函数。仿真结果表明,该方法具有较好的去噪效果。

Thresholding denoising method based on wavelet transform is an efficient method to reduce the white noise in the digital signal.Thresholding function has much importance to the continuity and precision of the reconstructed signal. However, hard thresholding method is discontinuous on the thresholding value, and soft thresholding method also has disadvantages. A new thresholding function is presented to overcome the shortcomings and it is the weighted average of the hard and soft thresholding function. Simulation results indicate that this new method has improved the effect of denoising.