An improved variational mode decomposition for seismic random noise attenuation using grasshopper optimization via shape dynamic time warping

Abstract

In seismic signal processing, attenuation of noise is one of the most difficult challenges. Variational mode decomposition (VMD) is a recently developed adaptable signal processing technique that subdivides a signal into bandlimited intrinsic mode functions (BIMFs). In VMD, the key elements are the number of intrinsic modes and the quadratic penalty factor for better denoising. In this paper, we have applied the Grasshopper optimization algorithm (GOA) to determine the key elements of VMD and decomposes the signal into optimized BIMFs. Then, the Shape dynamic time warping (shapeDTW) is applied for determining the relevant modes. Finally, the relevant modes are deal with Wavelet thresholding (WT) to obtain a better result of signal to noise ratio (SNR). Initially, the analysis is made on the synthetic signals to represent the dominance of the new approaching method with Empirical mode decomposition (EMD) based and VMD based methods. We also applied new approaching method to a 1D real seismic signal from the Pacific Earthquake Engineering Research Center (PEER) ground motion database to suppress the random noise interferences. The proposed approach is also tested in a 2D synthetic seismic section corrupted by random noise, and also with a 2D real Forrest seismic survey in South Australia's resources information gateway. Application of the proposed GOA-VMD-shapeDTW-WT method on both the synthetic and the real seismic signal shows supremacy with other denoising techniques.