Seafloor Mapping Research Articles

Feature fusion-based registration of satellite images to airborne LiDAR bathymetry in island area

Airborne laser bathymetry (ALB) and satellite imagery, as two different types of remote sensing (RS) data sources, have been widely used in coastal zone dynamic monitoring by virtue of unique advantages. However, ALB data and optical imagery collected in island areas generally have offsets due to various error sources, lack of distinctive features and scarce control. The feasibility and efficiency of existed models exist significant limitations, since these methods mainly extracts the form of prominent feature or pixel bias. To better serve policy makers, it is necessary to register heterogeneous data into a unified geographic reference frame. In this work, a feature fusion-based registration approach (FFBR) of satellite images to airborne LiDAR bathymetry is proposed for island change detection. It mainly attempts to accomplish the following two tasks: (a) maximize the utilization of features from heterogeneous data in the island area through underwater feature enhancement and feature fusion based on deep convolutional neural networks; (b) solve the mis-registration caused by the pixel difference source, by employing an improved active demons registration method that integrates cluster analysis, modal transformation and range constraints. Through comparative analysis with the SIFT-RANSAC and demons algorithm, the proposed approach is proved to have higher registration accuracy. Finally, the new model demonstrated the ability to detect changes in island cover and coastline.

Seabed Modelling by Means of Airborne Laser Bathymetry Data and Imbalanced Learning for Offshore Mapping.

An important problem associated with the aerial mapping of the seabed is the precise classification of point clouds characterizing the water surface, bottom, and bottom objects. This study aimed to improve the accuracy of classification by addressing the asymmetric amount of data representing these three groups. A total of 53 Synthetic Minority Oversampling Technique (SMOTE) algorithms were adjusted and evaluated to balance the amount of data. The prepared data set was used to train the Multi-Layer Perceptron (MLP) neural network used for classifying the point cloud. Data balancing contributed to significantly increasing the accuracy of classification. The best overall classification accuracy achieved varied from 95.8% to 97.0%, depending on the oversampling algorithm used, and was significantly better than the classification accuracy obtained for unbalanced data and data with downsampling (89.6% and 93.5%, respectively). Some of the algorithms allow for 10% increased detection of points on the objects compared to unbalanced data or data with simple downsampling. The results suggest that the use of selected oversampling algorithms can aid in improving the point cloud classification and making the airborne laser bathymetry technique more appropriate for seabed mapping.

Mapping of Seafloor and Delineating Coral Reefs of Southern Iraq/Northwest Arabian Gulf by Side Scan Sonar Technique

High-resolution Side Scan Sonar data, under-water observation (video and photograph) and bottom samples (rocks and sediments) were integrated to map the seafloor of the Palinurus shoal which is one of the submerged shoal site in the Iraqi territorial marine, Northwest region of the Arabian Gulf. The Side Scan Sonar field survey was carried out along with eight parallel profile extending in SE-NW direction, at a length of 2.8 km for each line. Four backscatter patterns were mapped within the study area. The pattern reflects various seafloor lithology. The average depth of sea bottom was determined at 7 - 10 m. Coral Reefs were mainly found at two parts of the shoal site in the formed assemblages and sporadic in central and northern parts of the site respectively. These coral take different shapes, sizes and colors. Also, they are live in some places and dead in others.

Communication-constrained cooperative bathymetric simultaneous localisation and mapping with efficient bathymetric data transmission method

AbstractBathymetric simultaneous localisation and mapping (SLAM) methods yield accurate navigation results for autonomous underwater vehicles (AUVs) and can construct consistent seabed terrain maps. Multiple independently working vehicles can complete tasks like surveying and mapping efficiently, which means cooperative bathymetric SLAM using multiple AUVs is suitable for large-scale seabed mapping. However, the transmission of bathymetric measurements collected using a multi-beam echo sounder over a low bandwidth, noisy, and unreliable acoustic channel is difficult, making cooperative bathymetric SLAM very challenging. This paper develops a graph-based cooperative bathymetric SLAM system that can compress many bathymetric measurements into small-scale acoustic packets and yield accurate navigation results with a 10% loss of acoustic packets caused by unreliable acoustic communication. According to the simulation conducted using the field data, the new algorithm is shown to be robust and capable of providing accurate location and mapping results over a low bandwidth, noisy, and unreliable acoustic channel.

Resolution dependence of acoustic scattering statistics for complex seafloors

Underwater Unmanned Vehicles (UUVs) utilize sonar perception to conduct sea floor mapping and target detection operations. However, cooperative surveys with multiple types of UUVs are difficult because platforms with different resolutions may generate different probability density functions (PDFs) of the magnitude of the complex pressure. An area of research that has not been adequately studied is the effects of resolution manipulation during the post-processing of high-resolution data from complex seafloor environments. This work analyzes synthetic aperture sonar (SAS) data collected from multiple seafloor geomorphologies surrounding Bergen, Norway to study the resolution dependence of scattering statistics for complex seafloors. Multi-look methods will be applied to reduce the resolution. The original data and reduced resolution data will be compared in terms of PDF amplitude and evaluated by standard goodness of fit tests with heavily tailed statistical models that are commonly used in the radar and sonar community, including mixture models. The goal of the paper is to provide a bridge to combining high-resolution and low-resolution sonar data together to enhance sonar perception. Data provided by the Norwegian Defense Research Establishment. [Funding provided by the Office of Naval Research.]

Applying a Multi-Method Framework to Analyze the Multispectral Acoustic Response of the Seafloor

Improvements to acoustic seafloor mapping systems have motivated novel marine geological and benthic biological research. Multibeam echosounders (MBES) have become a mainstream tool for acoustic remote sensing of the seabed. Recently, “multispectral” MBES backscatter, which is acquired at multiple operating frequencies, has been developed to characterize the seabed in greater detail, yet methods for the use of these data are still being explored. Here, we evaluate the potential for seabed discrimination using multispectral backscatter data within a multi-method framework. We present a novel MBES dataset acquired using four operating frequencies (170, 280, 400, and 700 kHz) near the Doce River mouth, situated on the eastern Brazilian continental shelf. Image-based and angular range analysis methods were applied to characterize the multifrequency response of the seabed. The large amount of information resulting from these methods complicates a manual seabed segmentation solution. The data were therefore summarized using a combination of dimensionality reduction and density-based clustering, enabling hierarchical spatial classification of the seabed with sparse ground-truth. This approach provided an effective solution to synthesizing these data spatially to identify two distinct acoustic seabed classes, with four subclasses within one of the broader classes, which corresponded closely with seafloor sediment samples collected at the site. The multispectral backscatter data also provided information in likely, unknown, sub-surface substrate differences at this site. The study demonstrates that the adoption of a multi-method framework combining image-based and angular range analysis methods with multispectral MBES data can offer significant advantages for seafloor characterization and mapping.

Remote Sensing of the Tautra Ridge: An Overview of the World’s Shallowest Cold-Water Coral Reefs

On the Tautra Ridge – a 39-100 m deep morainic sill located in the middle of the Trondheimsfjord, Norway – some of the world’s shallowest known occurrences of the scleractinian cold-water coral (CWC) Desmophyllum pertusum can be found. The earliest D. pertusum records from the Tautra Ridge date back to the 18th century, and since then, the location has provided easy access to physical coral specimens for numerous scientific studies. In 2013, the ridge was declared a marine protected area by the Norwegian Government due to its unique CWC reefs. However, few attempts have to our knowledge yet been made to characterize the distribution, extent and condition of these reefs extensively. The aim of the current study was therefore to add geospatial context to the Tautra CWC reef complex. In the study, data from multibeam echo sounding, synthetic aperture sonar imaging and underwater hyperspectral imaging are used to assess CWC reef occurrences from multiple perspectives. The study demonstrates how complementary remote sensing techniques can be used to increase knowledge generation during seafloor mapping efforts. Ultimately, predictive modeling based on seafloor geomorphometry is used to estimate both distribution and areal coverage of D. pertusum reefs along the majority of the Tautra Ridge. Our findings suggest that D. pertusum reef distribution on the Tautra Ridge is affected by several geomorphometric seafloor properties, and that the total reef extent in the area likely is close to 0.64 km2. Better description of current patterns across the Tautra Ridge will improve our understanding of the interaction between hydrography and geomorphology at the Tautra CWC reef complex in the future.

Mapping Underwater Bathymetry of a Shallow River from Satellite Multispectral Imagery

Rivers play an important role in water supply, waterway transport, and riverine species habitations. The underwater depth of a river channel is a fundamental geometric element and a key input to studies for the aforementioned and other applications. Traditional in-situ field surveys for flow depth measurements would incur high costs and encounter technical and/or logistical difficulties, particularly for river channels of large extents, very shallow channels, and remote, inaccessible channels. Recent advances in satellite remote sensing of inland water bodies like rivers, streams, and creeks have allowed mapping flow depth in an inexpensive, convenient, and efficient manner. The purpose of this paper is to demonstrate the retrieval of river flow depth from high-resolution (1.2 m) WorldView-3 satellite imagery. The depth retrieval methods are based on the ratio of top-of-atmosphere reflectance between two pairing wavelength bands of multispectral imagery. The originality of the methods lies in using analytical relationships without resorting to ground data of river flow depth for calibration, which improves from previous studies of remote sensing of river flow depth. The methods are successfully applied to mapping the underwater bathymetry of a 26 km reach of the Nicolet River in Quebec, Canada. This study shows the importance of geometric and radiometric corrections to the satellite images. The obtained flow depths using the ratio of reflectance of the red band (630–690 nm) to that of the green band (510–580 nm), among the eight bands in the visible spectrum, agree best with in-situ measurements. This study is perhaps the first use of the analytical approach for mapping river bathymetry. It is feasible to implement the approach to other river channels, with a good potential to reduce the costs and increase the efficiency of mapping river bathymetries.

Continuous, High-Resolution Mapping of Coastal Seafloor Sediment Distribution

Seafloor topography and grain size distribution are pivotal features in marine and coastal environments, able to influence benthic community structure and ecological processes at many spatial scales. Accordingly, there is a strong interest in multiple research disciplines to obtain seafloor geological and/or habitat maps. The aim of this study was to provide a novel, automatic and simple model to obtain high-resolution seafloor maps, using backscatter and bathymetric multibeam system data. For this purpose, we calibrated a linear regression model relating grain size distribution values, extracted from samples collected in a 16 km2 area near Bagnoli–Coroglio (southern Italy), against backscatter and depth-derived covariates. The linear model achieved excellent goodness-of-fit and predictive accuracy, yielding detailed, spatially explicit predictions of grain size. We also showed that a ground-truth sample size as large as 40% of that considered in this study was sufficient to calibrate analogous regression models in different areas. Regardless of some limitations (i.e., inability to predict rocky outcrops and/or seagrass meadows), our modeling approach proved to be a flexible tool whose main advantage is the rendering of a continuous map for sediment size, in lieu of categorical mapping approaches which usually report sharp boundaries or rely on a few sediment classes.

A first appraisal of the seismogenic and tsunamigenic potential of the largest fault systems in the westernmost Mediterranean

The westernmost Mediterranean hosts part of the plate boundary between the European and African tectonic plates. Based on the scattered instrumental seismicity, this boundary has been traditionally interpreted as a wide zone of diffuse deformation. However, recent seismic images and seafloor mapping studies support that most of the plate convergence may be accommodated in a few tectonic structures, rather than in a broad region. Historical earthquakes with magnitudes Mw > 6 and historical tsunamis support that the low-to-moderate instrumental seismicity might also have led to underestimation of the seismogenic and tsunamigenic potential of the area. We evaluate the largest active faults of the westernmost Mediterranean: the reverse Alboran Ridge, and the strike-slip Carboneras, Yusuf and Al-Idrissi fault systems. For the first time, we use a dense grid of modern seismic data to characterize the entire dimensions of the main fault systems, accurately describe the geometry of these structures and estimate their seismic source parameters. Tsunami scenarios have been tested based on 3D-surfaces and seismic source parameters, using both uniform and heterogeneous slip distributions. The comparison of our results with previous studies, based on limited information on the fault geometry and kinematics, indicates that accurate fault geometries and heterogeneous slip distributions are needed to properly assess the seismic and tsunamigenic potential in this area. Based on fault scaling relations, the four fault systems have a large seismogenic potential, being able to generate earthquakes with Mw > 7. The reverse Alboran Ridge Fault System has the largest tsunamigenic potential, being able to generate a tsunami wave amplitude greater than 3 m in front of the coasts of Southern Spain and Northern Africa.

Impact of bending-related faulting and oceanic-plate topography on slab hydration and intermediate-depth seismicity

Abstract It is commonly assumed that intermediate-depth seismicity is in some way linked to dehydration reactions inside subducting oceanic plates. Although there is growing evidence that the hydration state of an oceanic plate is controlled by its structure and degree of faulting, we do not have a quantitative understanding of this relationship. Double seismic zones offer the possibility of investigating changes in oceanic-plate hydration not only along strike but also with depth beneath the slab surface. To quantify the impact of oceanic-plate structure and faulting on slab hydration and intermediate-depth seismicity, with a focus on the genesis of double seismic zones, we correlate high-resolution earthquake catalogs and seafloor maps of ship-based bathymetry for the northern Chilean and Japan Trench subduction zones. The correlations show only a weak influence of oceanic-plate structure and faulting on seismicity on the upper plane of the double seismic zone, which may imply that hydration is limited by slow reaction kinetics at low temperatures 5–7 km below the seafloor and by the finite amount of exposed wall rock in the outer-rise region. These factors seem to limit hydration even if abundant water is available. Seismicity in the lower plane is, in contrast, substantially enhanced where deformation of the oceanic plate is high and distributed across intersecting faults. This likely leads to an increase in the volume of damaged wall rock around the faults, thereby promoting the circulation of water to mantle depths where serpentinization is faster due to elevated temperatures. Increased lower-plane seismicity around subducting oceanic features such as seamounts or fracture zones may also be caused by enhanced faulting around these features. Our results provide a possible explanation for the globally observed presence of rather homogeneous upper-plane seismicity in double seismic zones as well as for the commonly patchy and inhomogeneous distribution of lower-plane seismicity.

Development of a single-wavelength airborne bathymetric LiDAR: System design and data processing

Airborne laser bathymetry (ALB) is employed to measure shallow depth water by using a high sampling rate and point density. Thus, the problems of using traditional detection methods in inaccessible areas can be avoided. This study focuses on practical solutions for receiving echo signals, identifying target echoes, and integrating land and underwater terrain point cloud data in coastal environments. Optimization of the system design and its data processing scheme is undertaken to improve the performance of the receiving system based on a single-band ALB system developed by the authors at Shenzhen University. A flight experiment over eastern Hainan Island was conducted, during which the effectiveness of the proposed strategy was verified. Finally, the technical characteristics of the self-developed system are summarized to provide a reliable reference source for the subsequent industrialization and production of related marine light detection and ranging (LiDAR) laser systems.

A Simple Cloud-Native Spectral Transformation Method to Disentangle Optically Shallow and Deep Waters in Sentinel-2 Images

This study presents a novel method to identify optically deep water using purely spectral approaches. Optically deep waters, where the seabed is too deep for a bottom reflectance signal to be returned, is uninformative for seabed mapping. Furthermore, owing to the attenuation of light in the water column, submerged vegetation at deeper depths is easily confused with optically deep waters, thereby inducing misclassifications that reduce the accuracy of these seabed maps. While bathymetry data could mask out deeper areas, they are not always available or of sufficient spatial resolution for use. Without bathymetry data and based on the coastal aerosol blue green (1-2-3) bands of the Sentinel-2 imagery, this study investigates the use of band ratios and a false colour HSV transformation of both L1C and L2A images to separate optically deep and shallow waters across varying water quality over four tropical and temperate submerged sites: Tanzania, the Bahamas, the Caspian Sea (Kazakhstan) and the Wadden Sea (Denmark and Germany). Two supervised thresholds based on annotated reference data and an unsupervised Otsu threshold were applied. The band ratio group usually featured the best overall accuracies (OA), F1 scores and Matthews correlation coefficients, although the individual band combination might not perform consistently across different sites. Meanwhile, the saturation and hue band yielded close to best performance for the L1C and L2A images, featuring OA of up to 0.93 and 0.98, respectively, and a more consistent behaviour than the individual band ratios. Nonetheless, all these spectral methods are still susceptible to sunglint, the Sentinel-2 parallax effect, turbidity and water colour. Both supervised approaches performed similarly and were superior to the unsupervised Otsu’s method—the supervised methods featuring OA were usually over 0.70, while the unsupervised OA were usually under 0.80. In the absence of bathymetry data, this method could effectively remove optically deep water pixels in Sentinel-2 imagery and reduce the issue of dark pixel misclassification, thereby improving the benthic mapping of optically shallow waters and their seascapes.

AUV Navigation Correction Based on Automated Multibeam Tile Matching.

Ocean science and hydroacoustic seafloor mapping rely on accurate navigation underwater. By exploiting terrain information provided by a multibeam echosounder system, it is possible to significantly improve map quality. This article presents an algorithm capable of improving map quality and accuracy by aligning consecutive pings to tiles that are matched pairwise. A globally consistent solution is calculated from these matches. The proposed method has the potential to be used online in addition to other navigation solutions, but is mainly targeted for post processing. The algorithm was tested using different parameter settings on an AUV and a ship-based dataset. The ship-based dataset is publicly available as a benchmark. The original accurate navigation serving as a ground truth, alongside trajectories that include an artificial drift, are available. This allows quantitative comparisons between algorithms and parameter settings.

Assessment of PRISMA Level-2 Hyperspectral Imagery for Large Scale Satellite-Derived Bathymetry Retrieval

Currently, several satellite-derived bathymetry (SDB) workflows are based on a variety of satellite imagery which are analyzed by empirical or analytical methods. The latest availability of PRISMA hyperspectral data provides a new opportunity for testing their application in shallow water bathymetry mapping. Here we utilize two Level-2 PRISMA scenes from the Caribbean Sea capturing seafloor areas with diverse benthic features and we analyze them using the shallow water analytical models provided by the water-color simulator (WASI) software. The presented study examines the influence of spatial resolution and end-member spectra on the SDB output. Consequently, in one study area we apply inversion using additional reference spectra and in the other study area we exploit the PRISMA panchromatic band for producing a pan-sharpened, hyperspectral cube for bathymetry inversion. The results show good correlation with reference bathymetry data (sonar and admiralty chart) suggesting that PRISMA imagery has a clear potential in optical bathymetry studies. The use of appropriate end-member spectra assists in enhancing the accuracy of SDB, and pan-sharpened PRISMA imagery assists in improving the results when detailed bathymetry is required. PRISMA imagery can be effectively analyzed with open-source software WASI-2D and thus contribute new bathymetric data to regional-scale seafloor mapping projects.

Coralligenous banks along the western and northern continental shelf of Sardinia Island (Mediterranean Sea)

ABSTRACT Mapping of coralligenous banks was carried out along the continental shelf of the northern and western margin of Sardinia Island (Italy, western Mediterranean Sea) in the context of the European Marine Strategy Framework Directive (MSFD, 2008/56/EC). Coralligenous banks are bioconstructions produced by calcareous coralline algae. Seafloor mapping was carried out through multibeam echosounder surveys and video transects, using a Remote Operating Vehicles (ROV), in areas not formerly explored. A high-resolution digital model of the seabed (DTM) was obtained from multibeam data. A total surface of 436 km2 of sparse patches of coralligenous banks was mapped in the depth range ~40-160 m. A final map of coralligenous habitat distributions along the western and northern continental shelf of Sardinia (scale 1:250,000) was produced. The base-map is formed by the shaded DTM of the seabed. Other mapped features include the edge of the continental shelf and the distribution of rocky seabed.

Target Echo Detection Based on the Signal Conditional Random Field Model for Full-Waveform Airborne Laser Bathymetry

Airborne laser bathymetry (ALB) systems with digital full-waveform signal collection can obtain corresponding temporal positions from several backscattering surfaces by laser beam irradiation. This information can help describe the multi-elevation structures of the target and explore the echo signal attenuation response in different nonuniform mediums during laser propagation. Therefore, a full-waveform echo signal is quite practical for integrated water-land detection. However, the wavelength used in the ALB system is generally in the visible band range of 470~580 nm, and the received signal is constantly interfered with by many nontarget factors, such as imperfections in the receiving channel or the strong scattering from the transmission medium. The conventional processing method transforms nontarget interference into noise point cloud filtering or classification extraction, enabling the detection of a single surface or regular geometry. The accuracy of the identification and extraction for multi-elevation target surfaces echo signal is always reduced due to the significant noise signal intensity. We proposed a signal component detection method by constructing the echo signal feature functions and the conditional random field (CRF) model based on the full-waveform decomposition. The processing result for actual measurement data verified that the CRF strategy can effectively reduce the uncertainty of target surface detection. Compared with the single-beam echo sounder, the root mean square errors of the elevation deviation underwater were reduced by 3.2 cm and 4.9 cm respectively in the two different experimental areas.

Airborne LiDAR-assisted deep learning methodology for riparian land cover classification using aerial photographs and its application for flood modelling

Abstract In response to challenges in land cover classification (LCC), many researchers have experimented recently with classification methods based on artificial intelligence techniques. For LCC mapping of the vegetated Asahi River in Japan, the current study uses deep learning (DL)-based DeepLabV3+ module for image segmentation of aerial photographs. We modified the existing model by concatenating data on its resultant output port to access the airborne laser bathymetry (ALB) dataset, including voxel-based laser points and vegetation height (i.e. digital surface model data minus digital terrain model data). Findings revealed that the modified approach improved the accuracy of LCC greatly compared to our earlier unsupervised ALB-based method, with 25 and 35% improvement, respectively, in overall accuracy and the macro F1-score for November 2017 dataset (no–leaf condition). Finally, by estimating flow-resistance parameters in flood modelling using LCC mapping-derived data, we conclude that the upgraded DL methodology produces better fit between numerically analyzed and observed peak water levels.

Performance of deep-sea habitat suitability models assessed using independent data, and implications for use in area-based management

Marine spatial management requires accurate data on species and habitat distributions. For the deep sea, these data are lacking. Habitat suitability modelling offers a robust defensible means to fill data gaps, provided models are sufficiently reliable. We tested the performance of published models of 2 deep-sea habitat-forming taxa at low and high resolutions (~1 km and 200 m grid-cell size), across the extended exclusive economic zones of the UK and Ireland. We constructed new data-rich models and compared new and old estimates of the area of habitat protected, noting changes in the protected area network since 2015. Results of independent validation suggest that all published models perform worse than expected considering original cross-validation results, but model performance is still good or fair for Desmophyllum pertusum reef, with poorer performance for Pheronema carpenteri sponge models. High-resolution models using multibeam data out-perform low-resolution GEBCO-based models. Newly constructed models are good to excellent according to cross validation. New model spatial predictions reflect published models, but with a significant reduction in predicted extent. The current marine protected area network and the European Union ban on bottom trawling below 800 m protect 40 and 60% of D. pertusum reef-suitable habitat, respectively, and 11 and 100% of P. carpenteri-suitable habitat, respectively, within the model domain. We conclude that high-resolution models of D. pertusum reef distribution are a useful tool in spatial management. The poorer performing P. carpenteri model indicates areas for more detailed study. While low-resolution models can provide conservative estimates of percentage area-based conservation targets following the precautionary principle, high-resolution sea-floor mapping supports the development of better-performing models.

Cautious considerations for using multiple covariate distance sampling and seafloor terrain for improved estimates of rockfish density

Distance sampling is one of the most widely used methods for adjusting fish counts for detectability, which allows the estimation of absolute density. Further, abundance can be estimated spatially beyond the transect width with density surface models. The present study used a density surface modelling approach with survey data from line transects and terrain metrics from multibeam acoustic bathymetry data, integrated with a Geographic Information System, to estimate the population density of Pacific Ocean perchSebastes alutus(POP) within 6 sampling areas in the Gulf of Alaska (GOA) at 2 sampling scales: local and landscape. For adult POP, the final density model included depth as a significant predictor for density estimation at both the local and landscape scale. Additional factors included sponge coverage (local scale), aspect eastness and seafloor slope (landscape scale). Predicted densities of adult POP are highest on low (<5°) eastern-facing slopes with high sponge coverage (>75%), at depths of 100 to 200 m. By using habitat-based density models with data collected from line transect distance sampling and multibeam acoustic seafloor mapping surveys, we provide an alternative method for estimating rockfish abundance in untrawlable areas, which may help improve rockfish stock assessments and essential fish habitat descriptions and maps.