Underwater Acoustic Signal Denoising Algorithm Based On Improved EEMD And VMD

Underwater acoustic detection technology in the measurement of the underwater target,the respect such as development of Marine resources has been widely used,the hydrophone is one of the important equipment for underwater acoustic signal,but due to the complexity of the Marine environment,the underwater data will inevitably with a lot of noise and interference,the noise and interference makes useful signal cannot be accurately identified,resulting in the analysis of the underwater target deviation.In order to monitor,identify and orient the signal accurately,it is necessary to eliminate the influence of these noises.In this paper,the denoising methods of EEMD-NLMS-SSA,EEMD-WT-SSA and RSO-VMD-SSA are proposed on the basis of EEMD and VMD respectively,and verified by the measured signals of MEMS vector hydrophone.The main contents are as follows:1.After decomposing the signal by the collective empirical mode decomposition method,the decomposed IMF component is reconstructed by singular spectral analysis method,and then the desired signal is set,and the reconstructed signal is gradually approximated to the expected signal according to the recursive formula in the normalized minimum mean squared error algorithm(NLMS),which has obvious advantages in the results of the simulation experiment.2.EEMD was used to decompose the signal into a series of IMFs,and the high frequency signal and low frequency signal were distinguished by mean square error criterion.The high frequency signal was processed by wavelet threshold denoising(WT)method,and the processed high frequency signal and low frequency signal were reconstructed by singular spectrum analysis(SSA)method,and finally the denoising signal was obtained.The simulation experiment with different decibels and the actual lake experiment of North University of China show that this method can effectively remove the noise contained in the sound source signal,and has a good denoising effect.3.Rat swarm optimization algorithm(RSO)was used to optimize the penalty factor and decomposition layers in the variational modal decomposition algorithm(VMD),so that it could find the optimal parameters.The correlation coefficients between the original signal and each IMF component were calculated for the IMFs component obtained after VMD optimization,and the signals were divided into noisy components and noise components.Singular spectrum analysis(SSA)was used to reconstruct the noisy components,and the noise components were discarded.Simulation experiments and measured lake experiments show that the proposed method has obvious advantages in spectral graph,root mean square error and signal-to-noise ratio compared with other algorithms after adding different decibel noises.It can effectively remove noise and has certain practicability.Aiming at the problem of underwater acoustic signal denoising,several new joint denoising algorithms are proposed in this paper.According to the time domain frequency domain diagram and two evaluation indexes,the proposed algorithm has a good effect on noise removal and correction of baseline drift phenomenon,which lays a foundation for the subsequent analysis and processing of underwater acoustic signal.