Underwater acoustic target recognition has always played a pivotal role in ocean remote sensing. By analyzing and processing ship-radiated signals, it is possible to determine the type and nature of a target. Historically, traditional signal processing techniques have been employed for target recognition in underwater environments, which often exhibit limitations in accuracy and efficiency. In response to these limitations, the integration of artificial intelligence (AI) methods, particularly those leveraging machine learning and deep learning, has attracted increasing attention in recent years. Compared to traditional methods, these intelligent recognition techniques can autonomously, efficiently, and accurately identify underwater targets. This paper comprehensively reviews the contributions of intelligent techniques in underwater acoustic target recognition and outlines potential future directions, offering a forward-looking perspective on how ongoing advancements in AI can further revolutionize underwater acoustic target recognition in ocean remote sensing.

Deep learning has made data-driven methods effective in underwater acoustic target recognition (UATR) using passive sonar signals. However, a major current challenge is the limited availability of underwater acoustic data, leading to suboptimal performance without sufficient data. Self-supervised learning (SSL) can help address this problem by learning intrinsic patterns within acoustic data. Nonetheless, applying SSL in UATR systems requires efficient learning of meaningful representations that can provide quick prediction speed for real-time recognition systems. To this end, we propose the masking hierarchical tokens (MHT) method to learn meaningful representations via efficient self-supervised learning for our previously proposed UATR-Transformer, giving rise to the MHT-UATR-Transformer. In particular, the MHT-UATR-Transformer first exploits a new designed token-convolution-based hierarchical tokenization to efficiently obtain rich time–frequency information from the input Mel-spectrogram. Then, most of these tokens are masked with a high masking ratio and subsequently reconstructed by an integrated Encoder–Decoder structure. In this way, the MHT-UATR-Transformer can learn intrinsic representations of underwater acoustic signals to achieve better recognition performance with fewer labeled data, which is expected to alleviate the dependency on expensive acoustic data. Experimental results on two widely studied underwater databases show that our proposed method achieves better performance than supervised learning and state-of-the-art SSL method in both accuracy and speed, especially in few-shot and noisy scenarios, thus enhancing its practicality in real marine applications.

The accurate calculation of the sound field is one of the most concerning issues in hydroacoustics. The one-dimensional spectral method has been used to correctly solve simplified underwater acoustic propagation models, but it is difficult to solve actual ocean acoustic fields using this model due to its application conditions and approximation error. Therefore, it is necessary to develop a direct solution method for the two-dimensional Helmholtz equation of ocean acoustic propagation without using simplified models. Here, two commonly used spectral methods, Chebyshev–Galerkin and Chebyshev–collocation, are used to correctly solve the two-dimensional Helmholtz model equation. Since Chebyshev–collocation does not require harsh boundary conditions for the equation, it is then used to solve ocean acoustic propagation. The numerical calculation results are compared with analytical solutions to verify the correctness of the method. Compared with the mature Kraken program, the Chebyshev–collocation method exhibits higher numerical calculation accuracy. Therefore, the Chebyshev–collocation method can be used to directly solve the representative two-dimensional ocean acoustic propagation equation. Because there are no model constraints, the Chebyshev–collocation method has a wide range of applications and provides results with high accuracy, which is of great significance in the calculation of realistic ocean sound fields.

The accuracy and efficiency of sound field calculations highly concern issues of hydroacoustics. Recently, one-dimensional spectral methods have shown high-precision characteristics when solving the sound field but can solve only simplified models of underwater acoustic propagation, thus their application range is small. Therefore, it is necessary to directly calculate the two-dimensional Helmholtz equation of ocean acoustic propagation. Here, we use the Chebyshev–Galerkin and Chebyshev collocation methods to solve the two-dimensional Helmholtz model equation. Then, the Chebyshev collocation method is used to model ocean acoustic propagation because, unlike the Galerkin method, the collocation method does not need stringent boundary conditions. Compared with the mature Kraken program, the Chebyshev collocation method exhibits a higher numerical accuracy. However, the shortcoming of the collocation method is that the computational efficiency cannot satisfy the requirements of real-time applications due to the large number of calculations. Then, we implemented the parallel code of the collocation method, which could effectively improve calculation effectiveness.

Deep learning models can produce unstable results by introducing imperceptible perturbations that are difficult for humans to recognize. This can have a significant impact on the accuracy and security of deep learning applications due to their poorly understood interpretability. As a field critical to security research, this problem clearly exists in underwater acoustic target recognition for ocean sensing. To address this issue, this article investigates the reliability of state-of-the-art deep learning models by exploring adversarial attack methods that add small, exquisite perturbations on acoustic Mel-spectrograms to generate adversarial spectrograms. Experimental results based on real-world datasets reveal that these models can be forced to learn unexpected features when subjected to adversarial spectrograms, resulting in significant accuracy drops. Specifically, when employing the iterative attack method, the overall accuracy of all models experiences a significant decrease of approximately 70% for two datasets under stronger perturbations.

Geoacoustic inversion is a challenging task in marine research due to the complex environment and acoustic propagation mechanisms. With the rapid development of deep learning, various designs of neural networks have been proposed to solve this issue with satisfactory results. As a data-driven method, deep learning networks aim to approximate the inverse function of acoustic propagation by extracting knowledge from multiple replicas, outperforming conventional inversion methods. However, existing deep learning networks, mainly incorporating stacked convolution and fully connected neural networks, are simple and may neglect some meaningful information. To extend the network backbone for geoacoustic inversion, this paper proposes a transformer-based geoacoustic inversion model with additional frequency and sensor 2-D positional embedding to perceive more information from the acoustic input. The simulation experimental results indicate that our proposed model achieves comparable inversion results with the existing inversion networks, demonstrating its effectiveness in marine research.