一种改进的基于平移不变小波去噪方法及应用

An improved de-noising method and application based on translation invariant wavelet

传统的小波阈值去噪法会在信号的奇异点附近产生人为地振荡现象,影响了去噪效果.使用基于平移不变的小波去噪,可以很好地消除这种振荡现象,同时针对传统的软硬阈值函数的不足,研究了一种新的阈值函数,此阈值函数具有软硬阈值函数不可比拟的灵活性.将这种新的阈值函数与平移不变小波去噪法结合起来,提出了一种基于平移不变小波去噪的改进算法.最后通过理论仿真证明所提出的新方法在信噪比增益和最小均方误差意义上均优于平移不变小波去噪法以及传统的软、硬阈值去噪法.

The traditional wavelet threshold de-noising method may produce artifacts on the discontinuities of signal which has a bad effect on de-noising.Translation invariant wavelet de-noising method was used to eliminate this artifacts.And a new threshold function was used for the shortcomings of soft and hard threshold function.This new threshold function is more elastic than traditional threshold function.An improved de-noising arithmetic is put forward by combining the new threshold function and translation invariant wavelet de-noising method.Lastly simulation result indicated the new method gives better MSE performance and SNR gains than translation invariant wavelet de-noising and hard or soft threshold de-noising methods.