Adaptive ship-radiated noise recognition with learnable fine-grained wavelet transform. (1st December 2022)

• Record Type: Journal Article
• Authors: Xie, Yuan; Ren, Jiawei; Xu, Ji
• Is Part Of: Ocean engineering. Volume 265(2022)
• Publication Date: 2022-12-01
• Subjects: Ship-radiated noise recognition -- Adaptive generalized network -- Wavelet transform -- Parallel convolution attention -- Transfer learning; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.112626
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Analyzing the ocean acoustic environment is a tricky task. Background noise and variable channel transmission environment make it complicated to implement accurate ship-radiated noise recognition. Existing recognition systems are weak in addressing the variable underwater environment, thus leading to disappointing performance in practical application. In order to keep the recognition system robust in various underwater environments, this work proposes an adaptive generalized recognition system — AGNet (A daptive G eneralized Net work). By converting fixed wavelet parameters into fine-grained learnable parameters, AGNet learns the characteristics of underwater sound at different frequencies. Its flexible and fine-grained design is conducive to capturing more background acoustic information (e.g., background noise, underwater transmission channel). To utilize the implicit information in wavelet spectrograms, AGNet adopts the convolutional neural network with parallel convolution attention modules as the classifier. Experiments reveal that our AGNet outperforms all baseline methods on several underwater acoustic datasets, and AGNet could benefit more from transfer learning. Moreover, AGNet shows robust performance against various interference factors. Highlights: AGNet updates fine-grained wavelet-based parameters in a data-driven manner. AGNet applies the parallel convolution attention mechanism. AGNet shows a superior ability to benefit from prior knowledge by transfer learning. AGNet is robust against colored noise and low cut-off frequency.