摘要
针对未知含噪信号的小波去噪理论,提出了一种新的阈值方程,并基于斯坦恩无偏估计(SURE)优化算法和阈值方程寻找最优门限值,以确保去噪后的信号是对未知原信号的最优估计。同时,注意到使用正交小波去噪,容易在信号奇异点处产生Gibbs振荡。为解决该问题,在信号分解和重构时使用了平稳小波变换算法。最后,应用以上方法与基于SURE的正交小波去噪方法对2种含噪信号进行去噪分析,结果令人满意。
In this paper, according to wavelet denoising theory of signal, a new thresholding function is presented. And based on Stein＇s Unbiased Risk Estimate （SURE） optimizing algorithm and thresholding functions, we can find optimal threshold which makes the denoising signal he optimal estimation of original signal. At the same time, because of the Gibbs shake on singular point by orthogonality wavelet denoising, we use stationary wavelet transform to decompose and reconstruct signal. At last,we apply our method and orthogonality wavelet denoising method based on SURE to denoise two noise signals. The result is satisfying.
出处
《现代电子技术》
2006年第8期58-61,共4页
Modern Electronics Technique
