一种改进的EMD硬阈值去噪算法

An Improved EMD-based Hard Thresholding Denoising Algorithm

摘要:

为了提升基于经典小波阈值的EMD去噪算法的性能,利用高斯白噪声的统计特征提出了一种改进的硬阈值去噪算法；首先将含噪信号进行EMD分解,把第一个固有模态函数作为高频噪声直接去除并估算出其他 IMF中高斯白噪声的能量,然后根据硬阈值去噪的原理,利用滤除掉的样本点包含的能量等于白噪声的能量确定出合适的阈值；该方法能根据样本点自适应地确定阈值；最后通过对含噪正弦信号和仿真心电信号的去噪实验证实了改进后的阈值使算法去噪效果有明显提升。

Abstract:

An improved EMD-based hard thresholding denoising algorithm was proposed according to statistical characteristic of white Gaussian noise in order to increase performance of the EMD denoising algorithm based on classical wavelet threshold.Firstly, the noisy signal was decomposed by EMD, the first IMF was removed as high-frequency noise and noise energy included in the other IMFs was estimated.Sec-ondly, threshold parameter was determined based on hard threshold denoising principle and the criterion. that the total energy contained in the removed sample points is equal to the estimated white noise energy.The proposed method is self-adaptive. Finally,the improved method and the original method were imple-mented on noisy sinusoidal signal and artificial ECG signal.The results verify that the presented method can enhance denoising effect obviously.