A novel approach for underwater acoustic signal denoising based on improved time-variant filtered empirical mode decomposition and weighted fusion filtering

Denoising of underwater acoustic signal (UAS) has vital academic significance and practical value. To achieve effective denoising of UAS, a novel approach for UAS denoising based on improved time-variant filtered empirical mode decomposition and weighted fusion filtering is proposed. Firstly, to improve decomposition efficiency, time-variant filtered empirical mode decomposition (TVFEMD) based on an improved walrus optimization algorithm (IWaOA) (IWTVFEMD) is proposed. It decomposes signal into some intrinsic mode functions (IMFs), and IMFs are classified into high frequency IMFs and low frequency IMFs by energy analysis. Secondly, Gaussian-weighted moving average filtering (GWMAF) is used to filter boundary low frequency IMF and remaining low frequency IMFs are reconstructed as noise-free IMFs. Thirdly, all high frequency IMFs are reconstructed, and reconstructed high frequency IMFs are secondary decomposed by IWTVFEMD, and IMFs obtained by secondary decomposition are called SIMFs. K-means clustering and time-shift multi-scale amplitudeaware permutation entropy (TSMAAPE) are used to adaptively divide SIMFs into low complexity category and high complexity category. Then, weighted fusion filtering based on Gaussian mixture model clustering (GWFF) is proposed, which is used to filter low complexity category. High complexity category is discarded as noise. Finally, noise-free IMFs, boundary low frequency IMF after GWMAF and low complexity category after GWFF are reconstructed to obtain the denoised signal. Typical chaotic signal such as Chua and Duffing signal and actual measured five UAS are tested. The outcomes reveal that the proposed denoising method has achieved superior denoising results, which can improve signal-to-noise ratio of Chua and Duffing signal by 14 dB and 17 dB respectively, and make three-dimensional attractor phase diagram of actual measured UAS clearer and smoother.