Forward-looking sonar is one of the essential imaging equipment used in exploring underwater targets. However, it is always challenging to detect targets from sonar images considering the complex environment. This article represents an automatic underwater target detection method using clustering, segmentation, and feature discrimination. First, we combine the fuzzy C-means clustering (FCM) and K-means to cluster the sonar image globally to obtain as many regions of interests (ROIs) as possible. Second, the pulse coupled neural network (PCNN) is used to locally segment the target boundary from the ROIs. Finally, multiple features are extracted from the target area as the feature vector, which is inputted into the nonlinear converter to enlarge the features’ distance. Then we use Fisher discriminant to estimate the classification threshold, which realizes underwater target detection. The experimental results show that the proposed method has low detection error and good real-time performance under low false alarm probability, which is not inferior to popular deep learning approaches at present.

With the increasing demand for automatic underwater object detection, different deep learning (DL) methods have been performed on forward-looking sonar (FLS) images in recent years. However, DL-based approaches face challenges of overfitting due to excessive parameters and limited generalization ability. This study aims to develop a lightweight FLS detection network (FLSD-Net) that improves the detection of small underwater targets by addressing overfitting and generalization issues. The FLSD-Net incorporates three key modules - a light initial downsampling module (LID) to reduce information loss, a lightweight feature extraction module (LFE) to reduce overfitting from parameter redundancy, and an optimized network structure prioritizing high-resolution shallow layers to improve detection accuracy and speed. Validation on public and self-created datasets shows that the proposed method achieved a balance of detection speed and accuracy with better performance than state-of-the-art (SOTA) lightweight networks. In addition, the results in generalization experiments exhibit reduced missed and false detections, indicating improved robustness. This research advances underwater target detection by developing a lightweight network with improved generalization, providing a robust solution for real-world applications.

The increasing development of underwater recording and processing systems has created a demand for automated methods that accurately detect and track small underwater targets in sonar images. However, precisely tracking such targets in underwater environments remains challenging. This article proposes an improved automated underwater multitarget tracking method integrating threshold segmentation with the Gaussian mixture cardinalized probability hypothesis density (GM-CPHD) algorithm. Using the GM-CPHD algorithm instead of the Gaussian mixture probability hypothesis density (GM-PHD) helps mitigate estimation errors in target numbers, leading to improved accuracy in position estimation. In addition, to address the issue of missed detection, a secondary detection process is incorporated into the method, leading to a reduction in the missed detection rate. By considering the target area feature, the proposed method automatically sets an area threshold to distinguish between real targets and clutter, thereby reducing the average optimal subpattern assignment (OSPA) distance and minimizing the number of target trajectory breaks. The effectiveness of the proposed method is demonstrated through simulations and experimental tracking results. This method shows accurate tracking of multiple underwater targets, improved robustness, and holds potential for practical applications.

Moving target detection in the disturbing environment has been significantly active and challenging for underwater acoustic array signal processing. One of the most effective approaches is based on low-rank matrix estimation, which divides the reconstructed imaging frames into low-rank and sparse matrix components, with the low-rank structure representing the interference and the sparse component corresponding to the moving target. However, the array model has not been well explored, taking the reconstruction and moving target separation as independent steps. In this paper, the sparse Bayesian learning (SBL) approach is employed to recover the stable interference and the moving target simultaneously and separately, retaining the high reconstruction quality pertaining to SBL. Specifically, in each frame, a shared prior distribution is assigned to the stable interference, while the time-varying component (the moving target) is modeled with a unique prior distribution. The hierarchical model is solved by the variational Bayesian inference. The superior performance of the proposed algorithm is validated by simulation and experimental data in an underwater acoustic test tank and Songhua Lake in China, compared with state-of-the-art low-rank matrix estimation algorithms.

Direction of arrival (DOA) and source amplitude estimation are widely encountered in detection and communication as traditional array processing problems. Recently, sparse Bayesian learning (SBL) technique is increasingly exploited, which outperforms conventional compressive sensing (CS) approaches and provides an improved alternative for DOA and amplitude estimation. Although the accuracy of DOA is enhanced somewhat through SBL, the on-grid implementation of SBL suffers the grid mismatch errors due to predefined grids. The off-grid SBL can be employed to address the issue. However, the computation load of off-grid SBL is heavy because of per-iteration matrix inversion and slow convergence, which restricts its application for online systems. An off-grid space alternating multi-snapshot SBL (OGSA-MSBL) approach used for DOA and source amplitude estimation is proposed in this paper, which introduces the offset between the ground-truth DOA and the grid nodes to refine the model, and employs the space alternating (SA) approach to solve the problem efficiently. Compared with the original SA, the DOA accuracy can be improved even with a reduced number of grids, while the source amplitude can be estimated with a higher precision with the first-order Taylor refined beamforming model. As shown by the simulation and experimental results, the recovery accuracy of the proposed method is superior to the state-of-the-art on-grid SBL approaches and comparable to its off-gird counterpart, while much faster in terms of computation efficiency.

Existing direction of arrival (DOA) estimation approaches are often only considering Gaussian noise or impulsive noise, leading to the performance degradation in the scenario that both noises exist simultaneously. Considering that ambient noise of underwater acoustic array may have different variances due to the large aperture, this paper proposes a robust sparse recovery method based on variational Bayesian inference (VBI) that considers the “heavy-tailed” characteristics of impulsive noise, and the non-uniformity of ambient noise. Student-t distribution and Bernoulli distribution are modeled as impulsive noise in the measurement, and then, the array observed signal is created as a mixture of desired signal, impulsive noise and non-uniform noise. A VBI scheme is constructed to estimate the desired sparse signal to implement DOAs. Results obtained from the numerical simulation and experimental data processing verify the superior performance of the proposed VBI promoting DOA estimation for dealing with impulsive noise and non-uniform noise.