平移不变量小波去噪方法的研究

在小波去噪的方法中,应用最为广泛的是Donoho等人提出的非线性小波变换阈值法.但在有些情况下,如在信号的不连续点处,运用阈值法去噪会产生伪吉布斯(psuedo-Gibbs)现象.在阈值法基础上加以改进的平移不变量小波去噪方法不仅能有效的抑制伪吉布斯现象,而且能减小原始信号和估计信号之间的均方根误差(RMSE).通过仿真实验可以看出,该方法比阈值法有更好的去噪效果.

Study on translation invariant wavelet denoising method

In wavelet denoising methods,the mo st popularly used one is nonlinear wa velet transform thresholding method proposed by Don oho.But in some cases,such as in the n eighborhood of dis-continuities of signal,using thresholding denoising methods may exhib it psuedo -Gibbs phenomena.Translation Invariant wavelet denoising is an improvement based on thre sholding denoising method.It can not only supress psuedo -Gibbs phenomena,but also diminish RSME between original signal and estimated signal.Simulated exp eriment shows that this method is better than thresholding de-noising method.