Sonar Detection Research Articles

Cluster-driven non-uniform characteristic analysis of underwater target acoustic scattering field

Underwater small targets typically exhibit non-centrosymmetric geometries, resulting in a highly spatially inhomogeneous acoustic scattering field under active sonar detection. Addressing these challenges, this paper takes the hemispherical cylindrical shell as the research object, considers the angle continuity implied in the echo characteristics, and proposes a cluster-driven research method for the non-uniform characteristics of the target echo angles. First, the target echo features are extracted and feature vectors are constructed. Secondly, the t-distributed stochastic neighbor embedding algorithm is employed to improve the internal connection of the feature vector in the low-dimensional feature space and to construct the visualized feature space. Finally, the implicit angular relationship between echo features is extracted under unsupervised conditions by cluster analysis. The reconstructed local geometric structures corresponding to different categories demonstrate that the method effectively segments the angular intervals of local target structures based on their natural acoustic scattering characteristics. The study overcomes the inherent subjectivity of traditional methods for dividing angular intervals of target echoes, providing a more objective foundation for segmenting and analyzing the target’s geometrical structure.

Self-Training Can Reduce Detection False Alarm Rate of High-Resolution Imaging Sonar

Imaging sonar is a primary means of underwater detection, but it faces challenges of high false alarm rates in sonar image target detection due to factors such as reverberation, noise, and resolution. This paper proposes a method to improve the false alarm rate by self-training a deep learning detector on sonar images. Self-training automatically generates proxy classification tasks based on the sonar image target detection dataset, and pre-trains the deep learning detector through these proxy classification tasks to enhance its learning effectiveness of target and background features. This, in turn, improves the detector’s ability to distinguish between targets and backgrounds, thereby reducing the false alarm rate. For the first time, this paper conducts target detection experiments based on deep learning using high-resolution synthetic aperture sonar images at two frequencies. The results show that, under the conditions of equal or higher recall rates, this method can reduce the false alarm rate by 3.91% and 18.50% on 240 kHz and 450 kHz sonar images, respectively, compared to traditional transfer learning methods.

Optimized acoustic signal denoising using modified threshold shrinkage functions

Underwater noise is a significant issue that affects various underwater activities such as ocean exploration, submarine communication, SONAR detection, etc. De-noising stages are traditionally embedded in underwater activities; therefore, developing de-noising algorithms is a highly demanded field of study. In this paper, a suggested modification is applied to some well-known threshold functions that will be used in conjunction with complex wavelet transform (CWT). The modification is applied to the Garrote threshold function and two semi-soft threshold functions. CWT is used to decompose the signals, that are corrupted by measured underwater noise, then a modified wavelet shrinkage is applied to remove the noise and recover the original signal. The performance of the modified functions is compared with the original functions and the soft threshold function. The results demonstrate that the functions after the modification have better performance than the soft function and the original functions. The Garrote function has about 2.5 dB signal-to-noise ratio improvement and the semi-soft functions also have about 2 dB and 0.5 dB improvement respectively.

Improved YOLOv9 for underwater side scan sonar target detection

Improved YOLOv9 for underwater side scan sonar target detection

Elastic Spherical Cloak Driven by Bayesian Optimization

Abstract The introduction of machine learning algorithms has revolutionized the landscape of invisibility devices design, particularly in the intricate domain of elastodynamics. In this work, we proposed a core-shell configuration to realize elastic sphere cloaks, which was driven by Bayesian optimization algorithm to pinpoint the optimal configuration with high precision. Numerical simulations in solid and aqueous environments are implemented respectively to validate the cloaking efficacy of our design and the parameters identified by the algorithm. The results are in close alignment with theoretical prediction, underscoring the robustness of the proposed method. This approach provides new insights for the designing of elastic wave invisibility devices, and may lead to potential applications in underwater communication and sonar detection.

Lightweight global adaptive feature enhancement network for underwater object detection with sonar image

Abstract Sonar target detection is widely used in various underwater detection tasks. However, sonar images often lack target information and blurry features because of the interference of seabed environmental noise and complex background information. This poses significant challenges for sonar target detection tasks. A new lightweight Global Adaptive Feature Enhancement Network (GAFE-Net) is proposed to enhance the ability to acquire target information and effectively suppress background information. This network utilizes lightweight convolutional calculations and adaptive feature extraction blocks with global feature extraction capabilities to capture multi-scale semantic features of sonar images. Specifically, GhostConv is used to maintain local feature extraction capability while reducing computation complexity. The adaptive feature extraction block (C2FC) is employed to capture deep semantic features. The partial self-attention (PSA) mechanism is adopted to enhance the capability of focusing on targets. The Slim-neck is deployed for fusing multi-scale information. Validation on the public sonar image dataset URPC2021 shows that, compared to other advanced sonar target detection algorithms, the proposed method improves accuracy while maintaining low computational complexity, demonstrating excellent performance.

Effective deployment strategies for optimizing area coverage in multistatic sonar detection based on Cassini oval approximation and a virtual force algorithm

Multistatic sonar is regarded as a highly promising method for detecting quiet submarines over long distances. This technique exhibits high positioning accuracy, flexible configurations, easy concealment, and robust resistance to interference. However, the sources are more expensive than receivers and performance of multistatic sonar depends strongly on the relative positions of the sources and receivers. In this work, we study the problem of optimal deployment of receivers for one source and multiple receivers type multistatic sonar to achieve maximum area coverage. The area coverage of multistatic sonar is a multi-objective optimization problem (MOOP) balancing maximum coverage against minimal cost. To simplify the MOOP, the area enclosed by a Cassini oval is approximated to two circles and the normalized area difference is analyzed. Then the required number of receivers within a rectangular area of interest (AOI) is calculated based on full coverage theory. So the MOOP is simplified into a single-objective problem focused on achieving maximal coverage with a specific number of receiver nodes. An enhanced virtual force algorithm named Delaunay triangulation hole repair virtual force algorithm (DTHRVFA) is introduced to optimize the deployment positions of receivers as a heterogeneous nodes deployment problem. Simulation results show significant improvement in coverage compared with various methods of heterogeneous nodes deployment. The proposed method effectively addresses area coverage problems associated with the deployment of one source and multiple receivers type multistatic sonar receivers.

UUVDNet: An efficient unmanned underwater vehicle target detection network for multibeam forward-looking sonar

Precise localization of unmanned underwater vehicle (UUV) on multi-beam forward-looking sonar (MFLS) is a key technology in underwater robotic exploration. However, large appearance change and weak features of targets in MFLS image, pose a serious challenge on target detection. This paper proposes an efficient UUV detection model with an enhanced training scheme and convolutional block attention mechanism. The proposed training scheme is composed of dual-path gradient information optimization and adaptive loss adjustment, which improves the network’s detection capabilities and the underwater environment adaptability. The convolutional block attention module (CBAM) is developed for extracting channel and spatially weighted feature maps, and providing a mechanism to adaptive feature refinement. We construct a comprehensive sonar detection dataset from field experiments. By comparing existing state-of-the-art detection algorithms, our proposed UUVDNet outperforms Faster R-CNN, Cascade R-CNN, CenterNet, SSD, YOLOv8n, DPFIN, and MBSNN in terms of mAP50–95 by 39.5%, 35.7%, 8.4%, 14.0%, 2.5%, 13.6%, and 21.8%, respectively. Furthermore, our inference model size is the most compact among all detection models, with a size of 5.1M. Final ablation experiments effectively showcase the proposed components’ efficiency.

Underwater Image Recognition using Machine Learning

Machine Learning is the branch of Artificial Intelligence in which a computer is fed with data and based on that data it tries to find out solution on its own. It encompasses the procedure for feeding algorithms information to create the algorithms realize patterns in the data and then increase the performance of the algorithms. A Convolutional Neural Network (CNN) is a type of a deep learned an algorithm that has been created for image processing when using convolutional layers to automatically and in a hierarchical way learn features from the input images. Computers can perform well when it comes to image recognition and classification because of its capacity to detect and record such features as edges, or texture, and shapes among others. A rise in focusing on processing underwater images is essential for various research purposes necessary in marine biology, economy as well as in the management of species’ biodiversity. Observance of such organisms as plankton and Posidonia Oceanic allows determining environmental shifts, global warming, and impact of people on sea creatures. These include respectively planktons that are fundamental for oxygen generation, climatic events and the Posidonia Oceanic, which helps improve the sea Biodiversity and water quality. In the organisation study, image processing supplement the physio-chemical analysis and the sonar detection system. The performances of deep learning models, especially the CNNs, in underwater image processing are significantly better than the conventional methodologies. Preprocessing is important because images are often low-quality; data augmentation and transfer learning tackle the problems of a small dataset and class imbalance, which allow you to save computations during training. Through human activities, marine trash remains a menace to deep sea ecosystems and marine organisms calling for proper debris control.

Enhancing underwater target detection: Fusion of spatio‐temporal incompletely‐aligned AIS and sonar information via DTW and multi‐head attention mechanism

AbstractIn the field of underwater target detection, the passive sonar is an important means of long‐distance target detection. The sonar detection information typically includes both surface and underwater targets, whereas it is a great challenge on effectively distinguishing between surface and underwater targets solely based on sonar information. Effective fusion of sonar and AIS (Automatic Identification System) data can leverage their complementary nature to compensate for the limitation of sonar information. However, the sonar information and AIS information are acquired based on different detection principles and systems, which are essentially multi‐source heterogeneous information with obvious spatio‐temporal misalignment in nature. Existing fusion methods normally struggle to effectively align sonar and AIS data in both time and space subject to the complexity of the problem. In this study, the Dynamic Time Warping (DTW) algorithm is applied to align sonar and AIS data in the time domain. In addition, a deep learning algorithm with multi‐head attention mechanism is proposed to achieve the spatial alignment of sonar and AIS data, where the matching between the surface targets in AIS data and the same surface targets in sonar data can also be successfully achieved. It provides a priori knowledge to enhance the underwater target detection of the passive sonar by eliminating the interference of the surface targets. Based on the attention mechanism, the abstract features extracted from the intermediate‐layer of the neural networks are found to be effective to represent the typical features of the target motion trajectories, which also demonstrates the effectiveness of the attention mechanism. The experiment results show that the proposed method can successfully achieve a MatchingSucccessRate of over 95% between the AIS targets and sonar detection targets.

Effects of wall heating on wall pressure fluctuations and flow noise in a low-Reynolds-number turbulent channel flow with temperature-dependent viscosity

Wall pressure fluctuations and flow noise substantially degrade sonar detection performance and the acoustic stealth performance of underwater vehicles. This paper numerically investigates the effects of wall heating on wall pressure fluctuations in turbulent channel flow of water with temperature-dependent viscosity, exploring a novel method for controlling wall pressure fluctuations and flow noise in underwater vehicles. Large-eddy simulation (LES) is employed for the numerical calculation of the flow field, while a hybrid method combining LES with Lighthills acoustic analogy is employed to predict flow noise. The numerical results show that when the temperature difference between the wall and the incoming flow is 30 K and 50 K, the peak root-mean-square pressure fluctuations decrease by 6.76% and 8.91%, respectively. Wall heating stabilizes the pressure field near the wall, with the spectral levels of wall pressure fluctuations showing average decreases of approximately 1 dB and 2 dB. Wall heating weakens the energy-containing structures of wall pressure fluctuations and increases the overall convection velocity by 1.22% and 3.81%, respectively. Flow structure analysis reveals that the weakening of energy-containing structures results from the suppression of the vortex structures in the near-wall region. In the wall heating cases, peak turbulent kinetic energy decreases by 12.6% and 15.8%, respectively. Moreover, the sound pressure level of flow noise decreases with increasing wall temperature, with the maximum noise reduction exceeding 3 dB. Previous studies have not yet explored the effects of viscosity reduction caused by wall heating on wall pressure fluctuations and flow noise. This study demonstrates that wall heating is a promising method for reducing wall pressure fluctuations and flow noise.

The Correlation analysis between ambient noise level at low-frequency range and wind speed in shallow water

The ambient noise level in the ocean is an important consideration in the enhancement of passive sonar detection performance as well as undersea surveillance in a harbor area. The wind affects considerably the ambient noise level in the ocean. For this reason, many researchers have analyzed the wind dependency of ambient noise levels in the ocean in the last few decades. However, this analysis is mainly limited to more than the frequency of 100Hz, because the ambient noise level below 100 Hz has been affected by sea current. In this study, we have investigated ambient noise levels in the frequency range 10 Hz - 100Hz as the wind speed increases in shallow water. A drifting small vessel is used to exclude the effect of fluid noise generated on the surface of a hydrophone due to the sea current. The correlation between the ambient noise level and wind speed in frequency range 10 - 25 Hz showed from 0.77 to 0.95. Whereas, it was from 0.11 to 0.69 in the frequency range 32 -100 Hz. The maximum correlation appeared at 10 and 16 Hz. This was much higher than those measured in the frequency range 1kHz - 25 kHz.

A novel target detection method with dual‐domain multi‐frequency feature in side‐scan sonar images

AbstractSide‐scan sonar (SSS) detection is a key method in underwater environmental security and subsea resource development. However, many detection approaches primarily concentrate on tracking the evolution path of optical image object detection tasks when using acoustic images, resulting in complex structures and limited versatility. To tackle this issue, we introduce a pioneering dual‐domain multi‐frequency network (D2MFNet) meticulously crafted to harness the distinct characteristics of SSS image detection. In D2MFNet, a novel method for optimizing and improving the detection sensitivity in different frequency ranges called multi‐frequency combined attention mechanism (MFCAM) is proposed. This mechanism amplifies the relevance of dual‐domain features across different channels and spaces. Moreover, we introduce a dual‐domain feature pyramid network (D2FPN) significantly augments the depth and breadth of feature information in underwater small datasets. The methods offer plug‐and‐play functionality with substantial performance enhancements. Extensive experiments are conducted to validate the efficacy of the proposed techniques, and the results showcase their state‐of‐the‐art performance. MFCAM improves the mAP by 16.9% in the KLSG dataset and 15.5% in the SCTD dataset. The mAP of D2FPN was improved by 8.4% in the KLSG dataset and by 9.8% in the SCTD dataset. The code and models will be publicly available at https://dagshub.com/estrellaww00/D2MFNet.

Side-Scan Sonar Image Augmentation Method Based on CC-WGAN

The utilization of deep learning algorithms for side-scan sonar target detection is impeded by the restricted quantity and representativeness of side-scan sonar (SSS) samples. To address this issue, this paper proposes a method for image augmentation using a CC-WGAN network. First, the generator incorporates the Convolutional Block Attention Module (CBAM) to enhance the assimilation of global information and local features in the input images. This integration also improves stability and avoids mode collapse problems associated with the original Generative Adversarial Network. Subsequently, the CBAM is incorporated into the discriminator to facilitate a better understanding of the relevance and significance of input data, thereby enhancing the model’s generalization ability. Finally, based on this model, existing few-sample SSS images are augmented, and we utilize the augmented images for discrimination and detection with YOLOv5. The experimental results show that following training with the SSS dataset that is augmented by this network, the accuracy of target detection increased by 7.6%, validating the feasibility of our proposed method. This method presents a novel solution to the problem of low model accuracy in underwater target detection with side-scan sonar due to limited samples.

DNTFE-Net: Distant Neighboring-Temporal Feature Enhancement Network for side scan sonar small object detection

Side-scan sonar small target detection becomes fundamental work for side-scan sonar applications which holds vital importance for marine engineering, maritime military, and so on. However, existing algorithms suffer from performance degradation and poor portability when sonar data is limited. We find that benefit from the strong temporal correlation of sonar images, background information from the long-term images and foreground information from the short-term images can be used to enhance small targets in the current image. To fully utilize these two cues, this paper presents the Distant Neighboring-Temporal Feature Enhancement Network (DNTFE-Net) for detecting small objects in sonar images. The algorithm is based on the signal acquisition process of sonar devices, which is considered a continuous time sequence. Firstly, a flexible foreground feature alignment module is developed to align the foreground information between neighboring signals and the current signal so that foreground information can be used to enhance small target features. Then, we propose a background comparison module for selecting distant-temporal signals that provide the most contextual information in order to improve the ability of migrating to different domains. Finally, a foreground background feature amalgamation module is designed to fuse neighboring and distant-temporal signals for enhancing small target features. Furthermore, real sonar data is collected to compose the real continuous sonar dataset. Our DNTFE-Net exhibits superior recognition performance on the real sonar dataset compared to other popular methods, with an improvement of at least 27%. Our algorithms are publicly available at https://github.com/zbyhnu/DNTFE-Net.git.

Multi-scale fusion and efficient feature extraction for enhanced sonar image object detection

Sonar imaging is an underwater detection technology that relies on the transmission and reception of acoustic pulse waves. This technique plays a crucial role in various domains including underwater archaeology, energy exploration, and oceanographic surveying. A primary challenge in sonar imaging is the low signal-to-noise ratio and significant noise interference, issues that are influenced by the constraints of equipment performance and the underwater environment. Traditional object detection techniques have been inadequate in effectively extracting deep features from sonar images possessing targets with complex structures and have also demonstrated shortcomings in the fusion processing of target features at multiple scales, thereby affecting the robustness and accuracy of object detection. To improve the performance of sonar image object detection, we propose an advanced detection framework that integrates efficient feature extraction with multi-scale feature fusion. We employ EfficientNet as the backbone network. EfficientNet exhibits excellent feature extraction capabilities through comprehensive adjustments of depth, width, and resolution. We introduce a dual-channel attention module that blends Squeeze-and-Excitation (SE) and Efficient Channel Attention (ECA) mechanisms to amplify the expression of crucial feature channels and suppress the lesser ones. Additionally, we utilize a modified bidirectional feature pyramid network (BiFPN) to strengthen the integration of features across different layers. Employing these methods, we amalgamate the features into a shared weights classification network and bounding box prediction network for accurate target class discernment and localization. The experimental outcomes provide evidence of the notable superlative nature of the proposed framework in sonar image object detection, effectively ameliorating detection performance amidst noise interference.

Automatic Identification of Sunken Oil in Homogeneous Information Perturbed Environment through Fusion Image Enhancement with Convolutional Neural Network

With the development of the marine oil industry, leakage accidents are one of the most serious problems threatening maritime and national security. The spilt crude oil can float and sink in the water column, posing a serious long-term threat to the marine environment. High-frequency sonar detection is currently the most efficient method for identifying sunken oil. However, due to the complicated environment of the deep seabed and the interference of the sunken oil signals with homogeneous information, sonar detection data are usually difficult to interpret, resulting in low efficiency and a high failure rate. Previous works have focused on features designed by experts according to the detection environments and the identification of sunken oil targets regardless of the feature extraction step. To automatically identify sunken oil targets without a prior knowledge of the complex seabed conditions during the image acquisition process for sonar detection, a systematic framework is contrived for identifying sunken oil targets that combines image enhancement with a convolutional neural network (CNN) classifier for the final decision on sunken oil targets examined in this work. Case studies are conducted using datasets obtained from a sunken oil release experiment in an outdoor water basin. The experimental results show that (i) the method can effectively distinguish between the sunken oil target, the background, and the interference target; (ii) it achieved an identification accuracy of 83.33%, outperforming feature-based recognition systems, including SVM; and (iii) it provides important information about sunken oil such as the location of the leak, which is useful for decision-making in emergency response to oil spills at sea. This line of research offers a more robust and, above all, more objective option for the difficult task of automatically identifying sunken oils under complex seabed conditions.

A high-resolution DOA estimation via random forest virtual array extension

In sonar detection, underwater recognition, and similar fields, the challenge of using small-scale arrays is frequently encountered. With small-scale arrays, the degrees of freedom (DOFs) and accuracy of direction of arrival (DOA) estimation decrease significantly. Moreover, existing algorithms are less robust in complex environments and typically require substantial computing resources to ensure accurate estimation. To overcome these problems, this paper proposes a high-resolution DOA estimation via Random Forest virtual array extension (RFVAE). The algorithm first trains the random forest (RF) network using real array element data, and thus proposes a virtual covariance reconstruction technique. This technique allows for a substantial increase in the number of virtual array elements. Then, based on the virtual covariance reconstruction technique, a high resolution estimation network technique is proposed, which can increase the accuracy and DOFs of DOA estimation. Experimental results show that the proposed algorithm reduces error by 30% to 80% compared to recent technologies and can provide up to twice as many virtual elements. Additionally, the estimation speed can be increased by 3 to 1000 times, and it has significantly stronger robustness, functioning normally at signal-to-noise ratios 10 dB lower than the conditions under which other algorithms fail.

Multi-Beam Sonar Target Segmentation Algorithm Based on BS-Unet

Multi-beam sonar imaging detection technology is increasingly becoming the mainstream technology in fields such as hydraulic safety inspection and underwater target detection due to its ability to generate clearer images under low-visibility conditions. However, during the multi-beam sonar detection process, issues such as low image resolution and blurred imaging edges lead to decreased target segmentation accuracy. Traditional filtering methods for echo signals cannot effectively solve these problems. To address these challenges, this paper introduces, for the first time, a multi-beam sonar dataset against the background of simulated crack detection for dam safety. This dataset included simulated cracks detected by multi-beam sonar from various angles. The width of the cracks ranged from 3 cm to 9 cm, and the length ranged from 0.2 m to 1.5 m. In addition, this paper proposes a BS-UNet semantic segmentation algorithm. The Swin-UNet model incorporates a dual-layer routing attention mechanism to enhance the accuracy of sonar image detail segmentation. Furthermore, an online convolutional reparameterization structure was added to the output end of the model to improve the model’s capability to represent image features. Comparisons of the BS-UNet model with commonly used semantic segmentation models on the multi-beam sonar dataset consistently demonstrated the BS-UNet model’s superior performance, as it improved semantic segmentation evaluation metrics such as Precision and IoU by around 0.03 compared to the Swin-UNet model. In conclusion, BS-UNet can effectively be applied in multi-beam sonar image segmentation tasks.

Ultrathin and pressure-resistant meta-coating with embedded tree-shaped acoustic black hole for broadband low-frequency underwater sound absorption

Acoustic coating is the main technical means to absorb incoming sound from detection sonars. Considering the advance of sonar technology to cope with low frequencies and the underwater working environment, acoustic coatings are required to provide effective broadband low-frequency sound absorption (SA) under hydrostatic pressure. However, current underwater acoustic coating materials have poor low-frequency broadband SA performance, especially under hydrostatic pressure. Herein, we propose an ultrathin meta-coating with tree-shaped acoustic black hole (ABH) units embedded in an elastomer matrix. Each tree-shaped ABH unit consists of multi-scale ABH plates as scatterers, a center-support column as “trunk” and a frame as “skeleton”. This design combines the rich dynamics of multi-scale ABH plates, the pressure-resistance characteristics of the frame and the column, and the damping of the matrix, to collectively lead to superior performance. Capitalizing on the rich local and coupled large-amplitude local resonances of the unit, and strong resistance to static-pressure-induced deformation, high and quasi-perfect SA is achieved in multiple and ultra-wide low-frequency bands at a deep subwavelength scale under different hydrostatic pressures. Notably, quasi-perfect SA (>0.92) is achieved at 500 Hz with a meta-coating whose thickness is 2 % of the sound wavelength. The quasi-perfect SA is also demonstrated experimentally within 1200–7500 Hz, which persists till 1.0 MPa. The SA is still above 0.6 from 3–10 kHz at 4.5 MPa. For SA model under hydrostatic pressures, hyperelastic constitutive model and pressurized dynamic mechanical parameters of the matrix were established. Numerical results are then verified by experimental tests under different hydrostatic pressures. With both high local-modal densities and hydrostatic-pressure resistance, this design represents a new type of acoustic metamaterials and promises broad underwater applications.