Underwater acoustic signal analysis: preprocessing and classification by deep learning

Abstract

The identification and classification is important parts of the research in the field like underwater acoustic signal processing. Recently, deep learning technology has been utilized to achieve good performance in the underwater acoustic signal case. On the other side, there are still some problems should be solved. The first one is that it cannot achieve high accuracy by the dataset that is transformed into audio spectrum. The second one is that the accuracy of classification on the dataset is still low, so that, it cannot satisfy the real demand. To solve those problems, we firstly evaluated four popular spectrums (Audio Spectrum, Image Histogram, Demon and LOFAR) for data preprocessing and selected the best one that is suitable for the neural networks (LeNet, ALEXNET, VGG16). Then, among these methods, we modified a neural network(LeNet) to fit the dataset that is transformed by the spectrum to improve the classification accuracy. The experimental result shows that the accuracy of our method can achieve 97.22 %, which is higher than existing methods and it met the expected target of practical application.

Pre-crash control strategy of driver assistance system

“…It therefore differs from various modern approaches recently used e.g. in the Neural Network World journal [4,7,21,9,8,17].…”

Deep Learning-Based Classification of Raw Hydroacoustic Signal: A Review

“…The latest methods for underwater image enhancement are based on depth models and focus on ascertaining a mapping function from the subspace of underwater images to the subspace of ground truth images, but the use of these methods often leads to different background colors of underwater images as the diversity of underwater conditions are ignored. Wu et al (2020) [51] addressed the problem wherein assistance is needed to achieve higher accuracy after transforming a dataset into an audio spectrum. They used an improved neural network, LeNet, to fit the spectrally transformed dataset and achieved higher accuracy than the existing methods, thereby meeting the desired goal of being useful for practical applications.…”

A deep neural network approach for the prediction of protein subcellular localization

“…The authors Wen et al [11] designed a deep learning-based framework for the prediction of drug-target interaction. Further, the authors in [14] used deep neural networks such as LeNet, ALEXNET, and VGG16 for underwater acoustic signal processing. In writings, an embedding-based method is presented for predicting the subcellular localization of proteins [25].…”