摘要
经典的 D.L .Donoho小波阈值去噪算法在信号的奇异点处会产生 Pseudo Gibbs现象 ,从而使去噪后的信号在急剧变换部分产生振荡现象。为了抑制 Pseudo Gibbs现象 ,提出了一种基于平移不变的小波阈值去噪方法 ,在一定的平移量范围内 ,对分析信号进行循环平移 ,然后对这个新信号进行阈值处理 ,得到去噪后的信号 ,再对该信号作相反的循环平移 ,重复这一过程 ,对所获得结果求平均 ,就得到原始信号去噪后的估计信号。仿真实验的结果证明了算法的有效性。
Denoising algorithm of D.L.Donoho′s wavelet threshold may produce Pseudo Gibbs phenomenon on the singularity points of signal. To suppress the phenomenon, a denoising algorithm of wavelet threshold based on translationinvariance is presented. The algorithm performs the cyclespinning for the signal to be analyzed, and then, softthreshold method is used to shrink the wavelet coefficients of signal and reconstruct signal. This process can eliminate the oscillation on the singularities of signal. Simulation experiments show that the algorithm is available.
出处
《现代电子技术》
2003年第6期29-31,共3页
Modern Electronics Technique
