基于新阈值函数的小波阈值自适应去噪  

Adaptive wavelet threshold de-noising method based on a new threshold function

作　　者：姚宗娜 魏连鑫[1] 

机构地区：[1]上海理工大学理学院,上海200093

出　　处：《信息技术》2014年第11期12-15,20,共5页Information Technology

基　　金：国家自然科学基金(11301340);国家自然科学基金(11171041)

摘　　要：针对传统小波阈值去噪算法中阈值估计和阈值函数选取的不足,提出一种高阶可导的阈值函数,并利用牛顿迭代法求得Sure无偏估计意义下的最佳阈值,从而实现阈值估计和阈值函数相关联的小波阈值自适应去噪。仿真实验表明,新算法可获得比传统阈值去噪算法更理想的去噪效果。According to the defects of threshold estimation and threshold function selection in traditional wavelet threshold de-noising, a new threshold function is proposed that has high order derivative. Then the optimal threshold is calculated by Newton iterative method under SURE unbiased estimation. Thus, it sets up an adaptive wavelet threshold de-noising algorithm by connecting threshold estimation and threshold function selection. At last, the several simulation experiments are done and the results show that the new algorithm can obtain better de-noising effect than traditional methods.

关 键 词：小波变换 阈值函数 均方误差函数 自适应去噪 

分 类 号：TN911.4[电子电信—通信与信息系统]