Underwater Acoustic Interferometric Phase Denoising Method Based on Adaptive Multivariate Variational Mode Decomposition

Underwater acoustic interferometry is extensively employed in ocean observation due to its exceptional performance and high cost-effectiveness, serving as a critical tool for ocean development. Underwater elevation model reconstruction can be achieved efficiently and accurately by unwrapping the interferometric phase. However, the complex underwater environments often introduce noise into the interferometric phase, which affects the success rate of phase unwrapping and degrades the interferometry accuracy. In the process of mitigating interferometry phase noise, it is equally imperative to retain phase details that encapsulate nuanced variations in elevation. To address these challenges, an effective approach involves decomposing noisy interferometric phases into noiseless phase components and noise. This paper proposes an underwater acoustic interferometric phase denoising method based on adaptive multivariate variational mode decomposition. First, the sine-tent-cosine chaotic mapping and sine-cosine algorithm are combined with the subtraction-average-based optimizer, used to improve the optimization capability. Then, the improved subtraction-average-based optimizer is used to determine the optimal balance parameters and the number of modes for interferometric phase decomposition. Adaptive decomposition enhances the method’s adaptability to complex underwater environments. Lastly, an intrinsic mode function (IMF) evaluation parameter is utilized to distinguish IMFs and determine the signal subspace, while singular value decomposition is employed for noise reduction in noisy IMFs. Simulation and lake experiment results demonstrate that the proposed algorithm works well in all performance metrics. Our approach significantly reduces interferometric phase noise while preserving phase details, providing reliable interferometric phase unwrapping.