The lack of energy resources in shallow water depths made deepwater energy resources exploration as an important step. Deepwater energy devices like floating offshore wind turbines (FOWT), hybrid and integrated energy systems, and mobile offshore drilling units often need anchoring support, which is adaptable to all marine conditions. There are several marine anchor types like drag anchors, suction anchors and self-installing anchors, in which dynamically embedded anchors are popular due to the simple installation procedure. In the taut-leg or vertical mooring systems, these dynamically embedded anchor's performance is unsatisfactory. Also, their excess pore pressure dissipation time after installation in NC clay is higher, which made them lesser attractive. Despite its simplicity, this anchoring concept has not gained widespread acceptance for the above-mentioned reasons. Hence, the present work described the novel anchoring concept of foldable torpedo anchor (FOTOAN) and their working mechanism, theoretical capacity estimation and installation sequence in detail. A limited comparative laboratory study of the FOTOAN to the traditional torpedo anchor in soft clay shows the FOTOAN to be more efficient. In the vertical mooring systems, the performance is improved by 3.5 – 5 times. Thus, the FOTOAN has been promised as an efficient technique for deepwater anchoring applications. Also, the installation sequence of FOTOAN with the aid of two anchor handling vessels (AHV-1 & AHV-2) are presented as an example case study.
In recent years, demands for deep-sea utilization has been increasing from industries as well as natural science research, and the major utilization of the deep sea is being started soon. The authors have investigated the mix designs and construction methods of cementitious materials and structures for this trend. In this paper, our previous studies on the durability of cementitious materials under deep seas and in-situ casting tests of cementitious materials on deep seafloor are introduced.
Underwater images in global datasets are gathered by different cameras and under different lighting and altitude conditions. Image formation models can potentially compensate for these known differences and improve machine learning (ML) performance. In this paper we will investigate how ML trained on images from one system can classify images taken by another. We will train two ML classification models based on two different underwater camera systems. We are going to evaluate the performance of each model with data from both platforms and will demonstrate the improvement of using image formation models to make an image look-like it has been gathered with a different system and how ML performance is affected.
Digital Adaptive Optics enabled by homodyne encoding can mitigate oceanic turbulence in a passive imaging system. This effort demonstrates self-referencing homodyne interferometry that combines the passive imaging of multi-frame algorithm procedures with the single-frame correction capability of Shack-Hartmann adaptive optics techniques. By imaging QR Codes through a turbulent tank of clear water in the laboratory, we quantified the machine-readable performance gain provided by Digital Adaptive Optics when compared to a standard imaging camera. This work also presents temporal motion-compensation techniques to further improve image detection and pattern recognition. Results from this work verify that Digital Adaptive Optics with homodyne encoding provides ocean optical turbulence mitigation for single frames of data, paving the way for high-speed, self-contained imaging systems.
Multibeam echosounders are routinely used to provide information on the bathymetry and geomorphology of the seabed. At the same time, more and more users are exploring the full potential of sonar systems as a tool for marine environment monitoring. Benthic habitat mapping using remote sensing methods gives broad-scale information about the geographical range and distribution of marine habitats on the seafloor together with substrate characterization. The crucial element of data collection and processing for that purpose is backscatter, the echo intensity recorded concurrently with bottom detections. The acoustic response of the seabed depends not only on sediment type or its roughness but also on signal frequency and angle of incidence. By introducing more frequencies to our data collection we can capture more environmental information and make a prediction of the geographical distribution of benthic marine habitats and seabed types more accurate.
This paper presents the machine learning results to identify two similar targets in sonar images. We perform transfer learning with sonar images on YOLOv5 for similar images identification. Three YLOv5 pre-trained models with different neural network sizes were tested and compared.
The Japan Coast Guard (JCG) has been performing seafloor geodetic observation using the Global Navigation Satellite System-Acoustic ranging combination technique (GNSS-A), to reveal the subseafloor tectonic phenomena occurring at the subduction zones along the Japanese archipelago. JCG currently deploys 27 seafloor reference points named the Seafloor Geodetic Observation Array (SGO-A) along the Japan Trench and the Nankai Trough. Decadal GNSS-A observation at the SGO-A sites along the Japan Trench has revealed complex postseismic movements after the 2011 Tohokuoki Earthquake, indicating that afterslips may have occurred at the northern and southern ends of the coseismic slip, and that viscoelastic relaxation is continuing beneath the main rupture zone and also near the trench axis offshore of Fukushima. From the observations at the sites along the Nankai Trough, we have detected transient signals that may be caused by shallow slow slip events occurring near the trough axis. We are expecting further advancement in the observation frequency and positioning accuracy of GNSS-A to detect smaller, short-term subseafloor tectonic phenomena, to contribute to the study of megathrust earthquakes.
Analyses of marine ecosystems requires recording and observing large volumes of images and videos collected from various underwater platforms. In our work, we extensively use multi-camera systems mounted on Remotely Operated Vehicles (ROV) for obtaining videos of marine organisms. Along with these, ocean physico-chemical parameters such as temperature, salinity, oxygen concentration, vehicle position, etc. are also recorded. Analysis of videos and images is labor intensive and automated approaches are necessary to overcome this bottleneck. We use the annotation software Squidle+ for on-site and post-cruise annotations. Squidle+ allows combining images as well as corresponding metadata for annotations made by users and saves them it in a database. We developed an integrated system, built upon the Squidle+ annotation platform, that includes extensions to annotate and post-process multiple video data streams from ROVs with their associated metadata. Details of this SquidJam ecosystem are described in this paper. This ecosystem has allowed us to integrate and reduce time required for many tasks performed on the annotations to make scientific analyses.
The Global Navigation Satellite System-Acoustic ranging combination technique (GNSS-A) is a seafloor geodetic technique to provide underwater positioning information, which is affected by several kinds of positioning error sources. The uncertainty related to the sound speed structure (SSS), resulting from the temporal and spatial variation, significantly affects the positioning results. Specifically, the horizontal heterogeneity of SSS directly limits the horizontal positioning accuracy. In this paper, we use the equivalent gradient parameters in east-west and north-south components to depict the horizontal features of SSS and reconstruct the acoustic timing observation equation model. And then the extended Kalman filter (EKF) is applied for parameters estimation considering the superiority in dynamic measuring data processing. However, the filter accuracy is deteriorated and even divergent due to the pollution of unexpected measurement outliers. Therefore, the adaptively robust filter is presented and adopted to suppress noises effect in state-observation equation systems. To evaluate the performance of the above positioning methods, a series of simulation test were carried out. Testing results indicate that the positioning model considering of the horizontal heterogeneity of SSS performs better in extracting the synthetic time delays, and the adaptively robust filter shows advantage in resisting the effect of outliers and guarantees the positioning accuracy.
Dolphins use sound signal for their communication, feeding and localization. They use pulse train signals varying timing, intervals, range according to the application. The authors propose a utilization of dolphin like bio-mimic signals for acoustic localization by autonomous underwater vehicles (AUVs) in their task at the mobile surveillance system. The proposed imitating pulse train signal varies its amplitude and inter-click intervals randomly. Therefore, the detection error drastically decreases compared to that of single pulse signal. In simulations under real ocean environment, calculated received signal with biological ambient noise showed no detection error when the horizontal distance was under 200 meter using 101 repetition of pulse train. As the signal blends in with the environmental sound, it is possible to carry out missions undetected by intruders.
The last couple of years exacerbated the shortage of resources to conduct hydrographic surveys for dredging industry. At the same time the companies started to invest in USVs (Unmanned Surface Vehicles) or small SV (surface vehicles) to improve the mobilization time and reduce the overall "footprint" of the operation. In this paper, we present a paradigm change in managing the essential hydrographic surveyor resources by utilizing the technology to divide the dredging hydrographic survey into two parts, design and execution of the survey. The survey "design" is done by a hydrographer or survey manager operating from the remote location setting up the essential parts of the survey. Then the survey execution is done by a skipper or USV operator at the actual job site. These two are connected via a dedicated software which seamlessly combines these two functions into one consolidated efficient solution.
Large-amplitude internal solitary waves and internal tides are common oceanographic features in the South China Sea, which are generated in the Luzon Strait and propagate westward cross the basin area to the continental shelf and hit the east wage of DongSha Atoll. These waves shoal near Dongsha plateau and force nutrients and plankton and potentially entraining fishes, squids, and nutrients into the upper ocean, potentially foraged by cetaceans. Several studies have indicated that dolphins and whales may exploit prey aggregations associated with predictable but ephemeral ocean features, such as upwellings, fronts, and eddies. Prey availability for dolphins in the South China Sea influenced by the aggregative properties of internal waves is speculated and once was reported by one photo in 2005 in a benign sea state. This study presents evidence that internal waves attract and aggregate dolphins by using integrative analysis of current data, pressure data, and passive acoustic monitoring data.
Studying the amount, variety, and distribution of microscale particles and plankton at the global scale at different times and seasons is very important to understand ocean environments, and by reference, the global environment. It is important to image them in their natural habitats. Many optical techniques have been developed to meet this requirement. Most techniques suffer from one or another drawbacks, such as limited depth-of-field or low volume. Digital holography is an advanced optical imaging technique to image micro-objects, and it provides high-resolution recording, large depth-of-field and recording volume, and 3D viewing and tracking. This paper describes basic principles of in-line digital holography and provides common image processing methods. Four submersible digital holographic cameras, eHoloCam, RamaCam, weeHoloCam, and LISST-Holo, are introduced, as well as their image processing software. At the end, some limitations and challenges in the underwater holographic imaging systems are mentioned, and some possible solutions are discussed.
BioCam is a 4000 m depth rated high-resolution mapping instrument that uses lasers, strobes and cameras to generate multi-hectare 3D reconstructions of the seafloor at sub-centimetre resolution. These can be used to analyse seafloor ecology as well as the fine-scale features of seafloor terrains. BioCam was first deployed with the autonomous underwater vehicle (AUV) Autosub Long Range (ALR), also known as "Boaty McBoatface", in July 2022 using the research vessel RRS Discovery. During several dives a total of 80 ha of seafloor in the Greater Haig Fras and the South West Deeps (East) Marine Conservation Zones (MCZ) were visually mapped from altitudes between 4 and 5.5 m and sub-centimetre resolution bathymetry maps were generated. During the cruise, the AUV and BioCam were left onsite while the ship travelled to a new location, and both systems were controlled via satellite communication to upload new missions and confirm data quality, demonstrating the over-the-horizon operation capability needed to enable future ship-free deployments.
JAMSTEC has developed underwater optical wireless communication with a view to its use in underwater robots. Using a photomultiplier tube as a light receiving element and a wide light distribution LD as a light emitting element, a communication speed of 20 Mbps was achieved between underwater vehicles 120 m away. Furthermore, optical Wi-Fi is established between the autonomous underwater vehicle and the underwater station. Camera images in the station were transmitted when the autonomous underwater vehicle passes above the ocean lander through the optical Wi-Fi link. We now study underwater optical wireless communication device that transmits and receives optical signals at the end face of an optical fiber. It realizes arbitrary light distribution characteristics at arbitrary positions. In the tank test, the device was installed in the imitation underwater station. The underwater robot was operated in the tank wirelessly without a cable while we were watching the video image from the camera on the robot. In research on underwater image sensor communication using an underwater camera and an LED array type of underwater light, we control the blinking of each of the independent RGB LEDs arranged two-dimensionally on a plane, photograph them with a camera, and the data is decoded from the taken images. The optical axis alignment affects the communication performance greatly. We started to study on optical axis alignment (or optical tracking) using light itself. By combining rough tracking using a ring laser and fine tracking using a photon receiver, we try to align the laser optical axis with the opponent 300m away.
Underwater mapping is one of the necessary technologies for underwater exploration. However, when the sonar sensor, widely used to observe the surrounding environment, scans the scene, terrain’s height information can be lost. This paper proposes a method to restore the lost height information by adjusting the tilt of the sonar sensor with the pan-tilt mechanism. The height of the object can be calculated by analyzing the changes of the highlights of the object and the seabed change according to the sonar imaging model in the successive sonar images acquired while changing the tilt angle. The proposed method has been verified through a water tank experiments. The restored height information can be utilized for various underwater operations such as obstacle avoidance of autonomous underwater vehicle and underwater three-dimensional mapping.
The planet earth mainly consists of water, and monitoring and controlling pollution is essential. To ensure the surveillance and control of illegal or accidental spills of dangerous and harmful substances into the sea at the European level, the CleanSeaNet satellite monitoring system (service) was developed and implemented by the European Maritime Safety Agency (EMSA). This system’s main objective is to alert coastal states of possible spills in their Exclusive Economic Zone, thus allowing the necessary legal and combat actions (cleaning) to be taken in the affected area. A drift model for the spill behavior over time was proposed to be able to act on the right location, considering the existing weather conditions. In addition to verifying the areas most affected by spills and how they behave over time, considering the weather conditions, it is also necessary to propose a national network of unmanned vehicles (air and surface) that allows us to reduce costs and optimize monitoring and control tasks. It is shown that we could quickly decrease the response time and operational costs with the developed approach.
Semantic segmentation of marine images can be used to describe seafloor scenes and monitor marine creatures. However, preparing human-annotated datasets for image segmentation is time-consuming task. Therefore, this paper proposes a semi-supervised semantic segmentation algorithm based on the combination of Mean-Teacher and U-Net models to classify seafloor images collected in Philippines. The method will train and validate on two parts of the image. On the one hand, for images containing categories of coral, sea urchin, sea stars, and others (including sediment and seagrass), ordinary labeling is used for training and validation. On the other hand, for images only including seagrass and sediment categories, manual labeling of seagrass categories is particularly difficult. In order to overcome this barrier, based on the characteristics of this type of images, K-means clustering algorithm is used to obtain labeled dataset for training and validation. Compared with the U-Net based supervised method, the semi-supervised method proposed in this paper achieves good results and accuracy values even with fewer labeled images.
The importance of cobalt-rich crusts on the Pacific seamounts for possible future rare metal sources has been recognized these 40 years. The thin layer-type coverage and the micro-topographic undulation of substrates affect not only the excavation efficiency but also the economy of mining venture. Because of these distribution characteristics, no economic feasibility has been recognized through the past economic analyses. On the basis of a recent geological study about the basement rocks, a possibility of byproduct substrate recovery for phosphorous supply with cobalt-rich crust mining has been highlighted. In the mining model, the substrates are utilized as phosphorus ores for agricultural and chemical usages. Under some preliminary technical and economic assumptions, the possibility of combined mining of cobalt-rich crusts and phosphorous ores is examined on the basis of the byproduct usage. The results show a better economy of the venture including the substrate usage for phosphorous supply.
In mobile underwater acoustic (UWA) communication, such as communication between autonomous underwater and surface vehicles, the motion of terminals causes a time-varying phase shift to each multipath signal, degrading demodulation performance in a manner difficult to address through conventional signal processing. To suppress the effects of phase shifts, equalization with loops for the compensation of multipath signal phase shifts has previously been proposed. Although the improvement yielded by the proposed equalization has been confirmed through simulations in the previous study, the applicability of this approach to actual data in an at-sea experiment has never been reported. In this study, the proposed method's improvement yield was evaluated by application to acoustic data obtained in a prior marine experiment with an autonomous surface vehicle. Consequently, the equalization was demonstrated to improve demodulation performance in UWA channels with nonuniform Doppler shifts and multipath signals.
