Secondary decomposition multilevel denoising method of hydro-acoustic signal based on information gain fusion feature

Hydro-acoustic signal (HAS) processing is a hot research topic at present, which is of great significance to modern naval warfare, intelligent fishery management and marine ecological protection. However, HAS are inevitably polluted by marine environmental noise and noise from other equipment. In order to better denoise for HAS, secondary decomposition multilevel denoising method of HAS based on information gain fusion feature is proposed. Firstly, the optimized variational mode decomposition by chaos sine jumping spider optimization algorithm (CSJVMD) is proposed, and signal is decomposed by CSJVMD to obtain some intrinsic mode functions (IMFs). Secondly, information gain fusion feature (IGFF) is proposed, which divides IMFs into pure IMFs, mixed IMFs and noisy IMFs. Then, singular spectral decomposition (SSD) is used to secondary decompose mixed IMFs, and singular spectral components (SSCs) are obtained. SSCs are divided into pure SSCs, mixed SSCs and noisy SSCs by IGFF. Finally, improved non-local mean filtering (INLM) is proposed, and the mixed SSCs are adaptively filtered by INLM to reconstruct pure IMFs, pure SSCs and filtered mixed SSCs, so as to obtain the final denoised signal. The denoising experiments of simulated chaotic signal (SCHS) and measured HAS (MHAS) are carried out. The results show that it can increase the signal-to-noise ratio of SCHS by 7 similar to 14 dB, make the attractor phase diagram clearer and smoother, and MHAS noise is effectively suppressed. This lays a foundation for the subsequent processing of HAS such as prediction, detection, and feature extraction.