摘要
在讨论小波阈值去噪的软阙值、硬阈值方法基础上,提出了一种利用遗传算法对上下阈值T1和T2进行优化,并通过双阈值进行阈值量化的新方法,该方法不仅能有效地克服硬阈值处理方法可能引起的伪吉布斯现象和软阈值处理方法的恒定偏差的不足,还能有效地从高频信号中去除噪声引起的高频干扰信号。仿真实验结果表明:该方法在信号去噪中比传统的硬阈值和软阈值方法有更高的信噪比和更小的均方误差,可以较好地恢复信号。
On the. basis of discussing soft-thresholding and hard-thresholding methods for wavelet de-noising, a double-threshold is proposed and the genetic algorithm is used to optimize two thresholds T1 and T2. This new method can effectively overcome the Psuedo-Gibbs phenomenon of hard-thesholding method and intrinsic bias of soft-thresholding one. In addition, it can effectually remove high-frequency interfering signal arose out of noise from high-frequency signal. Simulated experiment results demonstrate that this new de-noising method can achieve higher signal to noise ratio (SNR) gain and lower mean square error(MSE) than traditional soft-thresholding and hard-thresholding methods in signal de-noising,and restores signal well.
出处
《传感器与微系统》
CSCD
北大核心
2007年第6期20-22,25,共4页
Transducer and Microsystem Technologies
关键词
小波去噪
双阈值
阈值规则
遗传算法
wavelet de-noising
double-threshold
threshold rule
genetic algorithm