Abstract—The Empirical Mode Decomposition (EMD) is a new method for processing nonlinear and non-stationary signals. In this paper, an improvement of a denoising scheme using EMD was proposed. A noisy signal was decomposed adaptively into IMFs by using EMD method. The basic principle of the original method was to filter by Savitzky-Golay filter (SG) each IMF and then to reconstruct the signal from the filtered IMFs. The new method for reducing noise was based on multiple pass the noisy IMFs through a Savitzky-Golay filter two or more times. The EMD-SG multiple pass denoising technique was tested on four noisy simulated signals (Doppler, Blocks, Bumps and Heavysine) corrupted by white Gaussian noise with different Signal-to-Noise Ratio (SNR). The performance of proposed denoising method was quantitatively evaluated using SNR and Mean Square Error (MSE). The evaluation results showed that the proposed EMD-SG multiple pass performed better than the original EMD-SG method.

 

Index Terms—EMD, denoising, Savitzky-Golay filter

Cite: Wahiba Mohguen and Raïs El’hadi Bekka, "Improvement of the EMD-SG Denoising Method," International Journal of Electronics and Electrical Engineering, Vol. 5, No. 1, pp. 26-29, February 2017. doi: 10.18178/ijeee.5.1.26-29