A profound understanding of unique underwater terrain characteristics is crucial for the Internet of Underwater Things (IoUT). Currently, using unmanned platforms equipped with Multibeam Echosounder Systems (MBES) for underwater topographic scanning has become a key method and future trend to understand seabed topography. However, non-ideal integration between MBES transducers and motion sensors induces dynamic integration errors (time delay, motion scale, yaw misalignment, lever arm errors), which manifest as high-frequency ship-track orthogonal bathymetric undulations, severely limiting the accurate generation of high-resolution seabed topographic maps. To address this, this study innovatively proposes a multibeam dynamic error detection method based on time-series neural networks (TSNN). Its core innovations are: first, leveraging the characteristic that dynamic errors’ ship-track orthogonal bathymetric undulations are highly correlated with attitude; second, focusing on distinct manifestations of beam-angle-dependent and beam-angle-independent dynamic errors across different beam reception angles, laying a key basis for accurate error inversion. The method involves two core modules: the "MBES Bathymetric Sequence Trend Extraction Module" separates seabed terrain trends from undulations to obtain error-driven fluctuation signals; the "Dynamic Error Regression Module" realizes quantitative prediction of dynamic integration errors. Experimental results show the method can effectively eliminate high-frequency undulations in shallow-to-medium water MBES data, with processing efficiency meeting unmanned platforms’ real-time requirements—providing a new technical approach for high-resolution marine surveying and mapping to support IoUT’s reliable deployment and operation.

To enable side-scan sonar (SSS) to be used for large-scale seabed sediment detection, classification, and composition analysis, this article breaks through the limitations of past superficial use of echo characteristics and proposes a hidden feature extraction method that combines imaging mechanisms with seabed reflection models. By using the penetration characteristics of sound waves, weak echo width (WEW) features were extracted from the water column region of SSS echoes. By using the Lambertian model and the concept of denoising diffusion probabilistic models (DDPMs), seabed topography and sediment reflection features were, furthermore, successfully decomposed from SSS echoes, and a method to extract features highly correlated with sediment composition from these components is also presented. The effectiveness of these hidden features was demonstrated through sediment classification and analysis experiments in the Bohai Bay of China. Compared to using only echo features, the hidden features extracted in this article improved classification accuracy by over 20% on common classifiers. Based on the hidden features and sediment distribution, the causes of seabed composition near Caofeidian were explained. Additionally, through feature importance and sediment response experiments, the response mechanisms of different hidden features to various sediment types were scientifically interpreted. This study demonstrates the significant application value of SSS in large-scale seabed sediment detection, classification, and analysis for geoscientific tasks.

There are common sound speed errors resulting from insufficient density of sound speed profile (SSP) stations in a multibeam survey. Existing methods typically use empirical orthogonal function (EOF) analysis to create new SSPs that can mitigate sounding errors coursed by sound speed errors, but they are often inefficient due to their iterative nature. To address this issue, we propose a method for constructing SSPs with surface sound speeds (SSS) constraints to reduce the sounding errors. This method begins by standardizing the measured SSPs through unified stratification, followed by the clustering of the standardized SSPs (SSSPs). Next, the clustered SSSPs, along with highly accurate SSSs from the transducer, are utilized to estimate the parameters for the inverse distance weighting (IDW)-based spatiotemporal interpolation formula. Finally, the determined formula is employed in the depth direction to construct an interpolated SSP (ISSP), which is then used to correct sounding errors. Experiments verified the proposed method and the results show that the root mean squares (RMS) of the sounding errors improved by approximately 83%, decreasing from 1.169 m of the alternative SSP (ASSP) to 0.202 m of the ISSP. Additionally, the computational time of the ISSP was reduced by a factor of 67 compared to the EOF method.

Backscatter image can be used to extract features for seafloor substrate classification. Currently, object-based method is a mainstream solution for image analysis. However, determining the optimal scale for image segment using traditional object-based method is a challenge. In this letter, we proposed an automatic method to select the optimal scale for each backscatter image pixel based on classification uncertainty. Then, the backscatter image features are extracted at the optimal scale and integrated with the bathymetric features to form the normalized multibeam echo sounder (MBES) features. Finally, a supervised classification model is established by relating the MBES features and ground truth data. Experimental results indicate that the overall accuracy reaches 85.41% and the kappa coefficient is 0.7935, which proved that the proposed method performed better than the traditional methods and can be transferable to other studies.

To overcome the challenges of limited samples, difficult acquisition, under-representation, and labeling in utilizing sonar images and deep learning for target detection, recognition, and segmentation tasks for full-class underwater targets, we propose the Seg2Sonar network based on SPADE. This network generates images through segmentation maps, thus eliminating the need for sample annotation. Additionally, we incorporate the Skip-Layer channel-wise Excitation (SLE) module into the SPADE network to enhance feature extraction ability with minimal training samples. To improve the realism of generated images, we introduce the Focal Frequency Loss (FFL) module, and propose the Elasticity loss (EL) strategy to improve the random combination capability of the network, considering the characteristics of low resolution and severe distortion of sonar images. Furthermore, we propose a weight adjustment (WA) strategy that tackles the challenge of low and unbalanced feature representation with few samples by taking into account the unbalanced distribution of features using prior information. hese four improvements enable efficient sample augmentation of sonar images with limited samples. Building upon the improved Seg2Sonar network, we propose an underwater full-class target augmentation strategy. Based on the imaging characteristics of sonar images, we classify underwater full-class targets into four categories: texture level, group level, shape level, and intensity level. We provide corresponding augmentation strategies by leveraging similar features among sonar target images or adding external radar/optical features to supplement the diversity of features. Our experimental results demonstrate the efficacy of our proposed method in achieving sample augmentation of underwater full-class targets with minimal samples (less than 10) or even zero samples. The approach achieves about 90% accuracy in detection, recognition, and segmentation for all types of targets through deep learning methods. Our findings provide a promising solution for efficient sample augmentation of underwater full-class targets with limited samples.

To meet the demands of marine scientific research and ocean engineering construction for cost-effective, high-resolution, and large-scale acquisition of seafloor topography, this study proposes an unsupervised terrain reconstruction technique for side-scan sonar (SSS). Unlike conventional methods, this technique does not rely on externally measured depths for initial terrain and ground truth; instead, it constructs these elements based on its own imaging mechanism. Through the development of a rigorous seafloor reflection model, we have designed an unsupervised depth convolutional neural network (DCNN) inversion framework. This framework utilizes the depth of the sea bottom line as the initial terrain and employs the height calculated from the target’s shadow as the ground truth. The proposed network not only extracts absolute topographical information but also captures relative changes in beam pattern and seabed substrate properties. The experimental results demonstrate the method’s excellent accuracy in terrain inversion for areas with water depths ranging from 10 to 14 m, achieving a vertical accuracy exceeding 0.17 m. Moreover, the method exhibits robust resilience to residual radiometric distortions, variations in seabed substrate, noise, and shadows. This innovative approach provides a reliable methodology for acquiring and broadly applying high-resolution seafloor terrain, potentially advancing various fields of marine research and engineering.

Matching multibeam echo sounder (MBES) data and side scan sonar (SSS) data has the potential to obtain high-accuracy seabed topography and geomorphological details simultaneously. However, the position uncertainty as well as radiometric and geometric distortion of SSS images bring serious challenges to the matching process. In this article, we proposed a novel matching method using sea bottom line matching and template matching to solve the above problems. First, we extracted the sea bottom points from the SSS image and calculated the depth series of the sea bottom line as the auxiliary information. Then, coarse matching is employed to this depth series and its corresponding terrain data of the MBES, to remove the coarse position uncertainty between the SSS and MBES images. After that, multiscale local self-similarity (MLSS) is proposed here, which is robust to geometric and radiometric distortion. Based on the coarse-matched SSS image, MLSS descriptor is applied for template matching between SSS and MBES images. Experiments verified the performance of the method. The high-resolution and high-accuracy seabed topography and surface details are eventually obtained.

In order to solve the problems of small samples, acquisition difficulties, under-representation and labeling difficulties in object detection, recognition and segmentation tasks for underwater all-category targets based on sonar images and deep learning methods. we propose a side-scan sonar full-class image sample augmentation method suitable for multi-task scenarios. Based on the superior image generation ability of the diffusion model, we use transfer learning to fine-tune the optical pre-trained model to build a side-scan sonar image generation model. Then, for the object detection task and semantic segmentation task, we use the image content and target shape as guidance information to guide the generation results of the diffusion model respectively. Meanwhile, proposed a mask synthesis method for SSS waterfall image generation based on the working principle of side-scan sonar. The synthesized mask images are used to guide the generation of side-scan sonar waterfall images. Finally, the underwater object detection and segmentation models are trained on the generated data. The experiment results show that training a model with generated data can be effective in improving accuracy.

Sub-bottom profiler (SBP) Sonar data are always polluted by noise that leads to the wrong interpretation. To deal with this situation, a method combining the prior information of SBP sonar data and the variational mode decomposition (VMD) framework is proposed attempting to recover clean data from its corrupted version. First, to deal with the unknown mode number of VMD method, a mode number determination criterion which takes the frequency distribution characteristic into consideration has been proposed. Then, the sea bottom line is extracted using the enhancement filtering and the threshold extraction method. The echoes above the sea bottom are from the water column and mainly contain noise components. To determine which modes are caused by the noise, we obtained and analyzed these echoes to achieve the noise prior frequency feature. After that, the noise modes can be determined and removed using its prior frequency information. Both simulated and real experiments have been used to test the effectiveness of the proposed method, and good performance has been achieved.

Sediment classification based on Chirp sonar data is very important in support of marine science and engineering. The traditional adopted spectral-ratio (SR) method is widely applied for its theory of simplicity and easy employment. However, the performance of SR method is heavily degraded by spectrum vibrations introduced by overlapping reflections. To solve this problem, a reflection signal decomposition method is proposed in this paper, which decomposes the overlapping reflection into separate reflection sub-signals to avoid the spectrum vibration. Firstly, we re-derive the overlapping reflection expression from the convolution model and give the initial model for the decomposition of the Chirp sonar signal. Then, introducing the spectrum smooth prior into consideration and incorporating it with the fidelity and the bandwidth terms, the decomposition model with constraints is proposed. After that, an iterative algorithm is introduced to solve the model optimization problem. The reflection sub-signal can be well obtained. Finally, through a least square linear fitting for the logarithm of the spectral ratio (the log-SR) of reflection sub-signals, quality factor (Q) which is applied for sediment classification can finally be achieved. Both real and simulated experiments have been conducted to verify the proposed method and some meaningful discussions are also drawn.

The use of side-scan sonar (SSS) images is limited by radiometric distortion, which is affected by factors such as angle response, beam pattern, and incident energy. Traditional methods have overlooked the impact of sediment-dependent angle response, degrading performance in complex SSS images. In this study, we propose a method for correcting radiometric distortion on SSS images based on the variational Retinex framework and sediment characteristics. First, we analyze the imaging theory of SSS and the angle responses of various sediments. We introduce a mixed Gaussian model to describe the feature of SSS backscatter strength (BS) distribution. After that, a radiometric distortion correction model, incorporating the sediment BS information constraint, low-rank constraint on the incident illumination component, and an anisotropic total variation (ATV) constraint on the albedo, is presented. BS information is obtained through sediment classification based on mixed Gaussian model optimization. A minimization optimization method is used to solve the model, and an SSS image with corrected radiometric distortion is obtained. Experiments with different SSS images demonstrate the effectiveness of the proposed method.

To address the drawbacks that current multibeam bathymetric outlier removal methods lack repeatability, often require parameter adjustment for different regions, still require a lot of manual labor, and offer limited scalability, a deep learning-based outlier removal method for multibeam bathymetric data is proposed. The method fully considers the multibeam data measurement principles and causes of outliers in multibeam bathymetric data and includes a comprehensive sample augmentation method and an outlier removal model based on a modification of a recently proposed PCPNet architecture. In our extensive evaluation, both on synthetic and real data, our method demonstrates robust outlier removal performance in a variety of marine environments without any parameter adjustment.

In the process of side-scan sonar image target detection and recognition, the direct application of deep learning techniques will cause serious overfitting due to the limited amount of sonar image sample data, thus restricting the accuracy of shipwreck detection and recognition. Therefore, based on the imaging mechanism and image characteristics of side-scan sonar (SSS), this paper proposes a joint ray model and sonar equation method for the simulation of SSS image samples. Primarily, the sonar equation is used to quantify the energy loss of the sound waves propagation underwater, and a ray model is established to simulate the transmission path of the sound waves underwater to realize the simulation of the SSS image. In addition, to further enhance the realism of the simulated SSS image, a target-to-target, background-to-background improved style transfer method is proposed and combined with a noise-added model to achieve the simulation of the impact of the marine environment on imaging. The method is applied to SSS target detection and identification, based on the SSS simulation imaging and data enhancement method to generate shipwreck samples for the training of the shipwreck detection model and testing with real data. The experimental results indicate that the accuracy of the model trained entirely on the simulation samples is comparable to the accuracy of model detection for training based on real samples, which verifies the reliability of the method.

In underwater perception and maritime surveys, due to the scarcity of training data and perturbation of speckle noise, the detection performance of underwater objects in side-scan sonar (SSS) images is limited. To address these problems, we proposed a noise feature disentanglement YOLO (NFD-YOLO) by combining noise-agnostic features learning and attention mechanism. Firstly, we rethink the speckle noise by treating it as the domain shift between the training dataset and real-measured SSS images and build a domain generalization-based (DG-based) underwater object detection framework. Then, we extend YOLOv5 with a feature manipulation module, a noise-agnostic subnetwork, and an auxiliary noise-biased subnetwork for noise features disentanglement, more biases toward noise-agnostic features and less reliance on noise-biased features in underwater object detection, respectively. Finally, the ACmix attention module is introduced for a more powerful learning capacity and attention to the object areas based on a small dataset. According to the experiment results, the proposed NFD-YOLO achieved 75.1% mean average precision (mAP) in the test domain, which increased by 7.5% than YOLOv5, and 75.7% ± 0.4% mAP and 77.5% ± 1.6% mAP for different speckle noise distributions and transfer directions, respectively, which verified its generalization ability and robustness for speckle noise. Therefore, the proposed method can mitigate the effects of speckle noise and provides a new thought to address the speckle noise in underwater object detection with a small dataset, which is of significance and benefits for underwater perception and maritime surveys.

Sound velocity errors often lead to distortion of seafloor topography in multibeam bathymetric data processing. The current method of sound velocity profile (SVP) inversion has the problems of small application scope of the inversed equivalent sound velocity profile (ESVP) and terrain distortion calculated by inversed SVP under the condition of water depth consistency constraint in overlapping line area. In this paper, we propose a method of SVP inversion using the Empirical Orthogonal Function (EOF) and Differential Evolution (DE) algorithm. Firstly, EOF is used to decompose the SVP matrix, then the water depth discrepancy between the check line and the main survey line is used as the objective function instead of the discrepancy of overlapping line area, and DE is used to find the optimal time coefficient, and finally, the SVP is reconstructed to correct the seabed topography. The method was verified by experiments in shallow and mid-deep water, and the mean absolute and standard deviation of the corrected bathymetric discrepancy values are about 50% of those before correction.

Air bubbles in seawater always lead to information loss for sub-bottom profiler (SBP) images. This common issue makes it difficult to interpret SBP data. Toward this end, this letter proposes a hybrid method to reconstruct the missing area on an SBP image. This method combines the information of multi-survey line to overcome the disadvantage of using a single survey line that has less observation information. First, based on an improved active contour model (ACM) algorithm and a bottom line match algorithm, the SBP images of adjacent survey lines are aligned, and the image patch group is obtained. Then, a multi-survey line patch group deep learning framework is trained to reconstruct the missing area using the patch group as the input layer. The global–local loss function is employed to optimize the deep learning training model. Both simulated and real experiments have been used to test the effectiveness of the proposed method, and good results have been achieved.

Reconstructing high-resolution underwater topography using side scan sonar (SSS) images with self-constraint may fail at the marginal part of the surveying area. To solve this problem, the bathymetric data are integrated as the reference to obtain high-resolution and high-accuracy topography. First, based on the seafloor reflection model, the SSS-based reconstructed terrain is obtained. Then, the integration model of SSS-based reconstructed results and bathymetric data in the same surveying area is proposed. Experiments have been used to test the effectiveness of the proposed method and good results have been achieved.

Wavelet threshold denoising with different threshold selection rules (TSRs) were used to reduce random noise (RN) in bathymetric laser full-waveforms. A nonreasonable threshold used for denoising can result in over-smoothing or under-smoothing of the signal and easily remove details of weak bottom return (BR). A unique and optimal TSR for all bathymetric full-waveforms of waters with different depths or turbidities is unavailable. Hence, an adaptive threshold selection (ATS) is proposed to improve the performance of RN reduction by adaptively selecting a threshold for each full-waveform based on the prominence of BR-to-noise ratio. The proposed method is applied to reduce the RN in raw green laser full-waveforms collected via Optech coastal zone mapping and imaging LIght Detection And Ranging (LiDAR). Compared with other traditional methods, the ATS improves the ratio of detectable BR by 5.64% and achieves a root mean squared error (RMSE) closer to the real RN level. Therefore, ATS can effectively remove the RN, enhance the prominence, and ensure the fidelity of weak BR.

To solve the shortage of training samples when using deep learning to detect shipwrecks, a comprehensive sample augmentation method is proposed. The method fully considers the imaging mechanism and environment of side-scan sonar (SSS), such as acoustic emission and reception, waterbody, target reflection, and seafloor background, and generates diverse and representative shipwreck samples from five aspects of target diversity, target texture, imaging resolution, equipment and environmental noises, and background through a series of novel sample augmentation methods. Under the condition of zero real SSS samples, a detection model of YOLOv5s was established with these amplified samples and achieved a mean average precision (MAP) better than 96% for real SSS data detection.

For faster and more accurate acquisition of the spatial incidence angle of multibeam echosounders, a better algorithm called “Cross” is proposed in this paper, where the spatial incidence angle can be calculated by solving the simultaneous equations of the transmit and receive cone equations. In the algorithm, the two cone equations for the transmit and receive sectors are established using 14 known angles. Then the intersection vector of both cones at the same apex can be calculated very quickly from the simultaneous equations of transmit and receive sectors, and the spatial incidence angle and relative azimuth of each beam can also be obtained from them. Comparison with other algorithms, calculating the spatial incidence angle is a generic method regardless of shallow or deep water. Furthermore, “Cross” algorithm achieves a similar internal precision and about 12.3% - 13.4% speedup than “UNB”. The effectiveness and practicability of the proposed algorithm are verified through the experiments of various instrument models and file formats.

Radiometric distortion caused by the time-varying gain (TVG), beam patterns, angular responses, and sonar altitude variations, highly degrades the quality of side-scan sonar (SSS) images. Thus, radiometric distortion correction becomes a fundamental step for SSS image processing which holds vital importance for geomorphic applications. However, existing methods cannot take the prior information of the acoustic illumination component as well as the feature of seafloor into consideration well, which would easily cause damage to the image and also always be powerless for residual stripe noise. In this paper, a novel radiometric correction method is proposed. First, we give a detailed analysis of the SSS imaging theory based on the Lambert’s law as well as the prior knowledge about the characteristics of SSS images. Then, incorporating the prior of the SSS imaging process, the low-rank constraint is specifically introduced for the illumination component, while the anisotropic total variation (ATV) constraint is used to constraint the albedo component, combining other constraints, a decomposition model is proposed to correct the radiometric distortion based on the SSS imaging theory. And an alternative minimization method has been adopted to solve the proposed model effectively. Experiments proved the validity of the proposed method.

Nowadays, a 3-D sub-bottom profiler (SBP) can produce the point clouds of the subseabed and is gradually receiving more attention in getting geologically significant surfaces to reveal sedimentary environments and structural features. However, little literature studied the automatic extraction of these surfaces from the 3-D SBP data currently. Thus, this article proposes a hybrid method consisting of a surface extraction algorithm and a segmentation algorithm. First, the multiprofile SBP data are converted into 3-D data volume. Second, by taking full advantage of the plate-like characteristic of the layer surface in the 3-D SBP data, a plate-like enhancement filtering algorithm based on the nonuniform Gaussian scale is given to filter the 3-D data volume. Third, a threshold extraction is applied to extract surface voxels, and a hybrid region growing algorithm is put forward to segment surface voxels into basic units by combining multicriteria. Finally, the surface segmentation problem is formulated as global energy optimization, and a stepwise segmentation algorithm is proposed to get the final surface set. To verify the effectiveness of the proposed method, experiments were conducted and analyzed. The results showed that the proposed method performed well.

The presence of sound velocity errors (SVEs) is common in multibeam sounding. Existing methods are good at eliminating these errors in the area with flatter seabed terrains but are not suitable for deep and complex seabed terrains. Therefore, a universal method is proposed in this paper to eliminate these errors in various depth and terrain areas with a regression model and the equivalent sound speed profile (ESSP). The vertical biases are calculated at these homonymy points of the main survey line and the inner swaths of the auxiliary survey line firstly. Then, a regression model is built by associating the vertical bias with the initial incident angle and travel time, and the regression coefficients are estimated with the help of the partial least squares (PLS) method. On the basis of the determined regression model and ESSP, the effects of sound errors for multibeam sounding data are eliminated. The proposed method was verified by the multibeam sounding data measured in the medium-shallow and deep water areas, and the standard deviations of the absolute values of depth differences (AVDDs) after correction were both about 2 times lower than those before correction.

The towing operation mode of side‐scan sonar system (SSS) easily results in dislocations and distortions of targets in the SSS strip image, brings difficulty to the mosaic of these strip images by the geocoding method or the tessellation‐line method and affects the recognition and understanding for seabed relief. Therefore this study proposes a new method, namely the segment‐image mosaic method based on corresponding features of the two adjacent SSS strip images. Through SSS image preprocessing, segment match based on corresponding features and image fusion in the common coverage area based on wavelet transformation, this new method overcomes the drawbacks of the traditional methods of image mosaic, fulfils well the mosaic of SSS strip images and finally a whole‐area SSS image is formed. Experiments have verified that the mosaic image formed by the new method can correctly reflect the position, shape and distribution of seabed targets, which is helpful to understand seabed relief.