Denoising of Lightning Electric Field Signals Based on EMD-Wavelet Method

Abstract

In this paper, a new lightning electric field (LEF) signals denoising approach combined with empirical mode decomposition (EMD) and wavelet transform (WT) is proposed. Unlike the conventional EMD or WT denoising approaches, we use EMD to decompose the LEF signal firstly, the continuous mean square error(CMSE) criteria are used to determine a turning point in the original signal energy, then the Birge-Massart threshold wavelet denoising method is employed to denoise the high frequency component which contains lots of noise. Finally, the clean high frequency component and the remaining low frequency intrinsic mode function, and the residual of the EMD operation are employed to synthesize a cleaner LEF signal. The method is illustrated on real data, and the performance of the proposed method is evaluated in terms of several standard metrics. The results show that the proposed method is able to reduce noise from the noisy LEF signals more accurately and effectively in comparison to EMD filtering and Wavelet filtering methods.