基于一种新的指数收缩函数的小波去噪方法

摘要

首先介绍一些基于小波去噪的经典方法,主要集中在小波收缩函数上,然后提出了一种新的柔性收缩函数模型,相比以前的收缩函数,这种函数具有较好的柔性结构,能够较好地提高去噪能力.新的收缩函数避免了硬收缩函数所带来的不连续性和软收缩模型带来的偏差估计.利用Matlab进行了仿真实验,用仿真实例说明了这种指数收缩函数相比以前的函数能够提供更高的SNR增益和更小的MSE性能.%A new exponential shrinkage function is proposed, which has a flexible structure compared with the shrinkage methods proposed previously. The new shrinkage function solved the problem of function encountered by using a hard shrinkage function and will also alleviate the bias in estimation caused by using a soft shrinkage function. The formulae for calculating the bias, variance and L2 risk of the exponential shrinkage function are derived. Simulation examples are given to demonstrate the application procedure of the new method. It is clearly shown that the proposed exponential shrinkage method gives higher signal noise ratio (SNR) gains and reduces the mean square error (MSE) values in evaluation of the signal demising performance.