Ship-Radiated Noise Separation in Underwater Acoustic Environments Using a Deep Time-Domain Network

Qunyi He,Haitao Wang,Xiangyang Zeng,Anqi Jin

DOI: https://doi.org/10.3390/jmse12060885

IF: 2.744

2024-05-27

Journal of Marine Science and Engineering

Abstract:Ship-radiated noise separation is critical in both military and economic domains. However, due to the complex underwater environments with multiple noise sources and reverberation, separating ship-radiated noise poses a significant challenge. Traditionally, underwater acoustic signal separation has employed blind source separation methods based on independent component analysis. Recently, the separation of underwater acoustic signals has been approached as a deep learning problem. This involves learning the features of ship-radiated noise from training data. This paper introduces a deep time-domain network for ship-radiated noise separation by leveraging the power of parallel dilated convolution and group convolution. The separation layer employs parallel dilated convolution operations with varying expansion factors to better extract low-frequency features from the signal envelope while preserving detailed information. Additionally, we use group convolution to reduce the expansion of network size caused by parallel convolution operations, enabling the network to maintain a smaller size and computational complexity while achieving good separation performance. The proposed approach is demonstrated to be superior to the other common networks in the DeepShip dataset through comprehensive comparisons.

oceanography,engineering, marine, ocean

What problem does this paper attempt to address?

### The Problem the Paper Attempts to Solve

This paper aims to address the issue of separating ship radiated noise in underwater environments. Specifically, due to the complexity of underwater environments, which include multiple noise sources and reverberation, separating ship radiated noise is a significant challenge. Traditionally, underwater acoustic signal separation has often employed blind source separation methods based on Independent Component Analysis (ICA). However, in recent years, this problem has begun to be viewed as a deep learning problem, where the characteristics of ship radiated noise are learned from training data for separation purposes.

The paper proposes a deep time-domain network based on parallel dilated convolutions and parallel grouped convolutions to achieve the separation of ship radiated noise. This method uses parallel dilated convolution operations with different dilation factors to better extract low-frequency features in the signal envelope while preserving detailed information. Additionally, grouped convolutions are utilized to reduce the network size expansion caused by parallel convolution operations, thereby maintaining a smaller network size and computational complexity while achieving good separation performance.

Comprehensive comparisons on the DeepShip dataset demonstrate that the proposed method outperforms other common networks.