Cost-effective ecological monitoring in shallow waters using amphibious unmanned aerial vehicles (AUAV) and deep learning-based computer vision.

An automatic active contour method for sea cucumber segmentation in natural underwater environments

In various inspections for the construction and the maintenance of port facilities, the underwater observation is a very important and an effective method. In general, the view of underwater video images is narrowly limited due to the turbidity of the sea water and the camera field angle. Therefore, the data processing such as mosaicing is necessary to observe the object over the wide range. However, the automization of mosaicing work is difficult, because of the insufficiency of the position information about underwater images. We have proposed an efficient mosaicing system of underwater images using AUV (autonomous underwater vehicle) mounting the video camera with the line-laser. It is preferable to take a movie from the same direction by a constant distance as much as possible to make an accurate mosaicing image efficiently. In proposed system, the attitude of AUV is controlled based on the distance and the inclination to the object calculated from the position of the laser line in obtained images. To make mosaicing image by proposed method, the image of the port structure with the laser line is acquired in a real sea area. And, the affine transformation, the projective transformation, and the method of using information about the laser line are used as a transformation method of images extracted from a movie of an actual port structure. In this paper, applicability and the improvement method of mosaicing procedure are discussed by comparing mosaicing images obtained by these methods.

A manipulative field experiment was designed to test the hypothesis that deposit feeding by an epibenthic holothurian, Parastichopus parvimensis (Clark 1913), influences the distribution and abundance of infaunal populations. The experimental design allowed separation of the effects of P. parvimensis from the other common bottom disturbers in the area, the labrid fishes Semicossyphus pulcher and Halichoeres semicinctus. Five "treatments" were replicated twice: 1) complete exclusion - no fish or sea cucumbers, 2) partial exclusion-fish allowed in, sea cucumbers excluded, 3) sea cucumber enclosure -1 sea cucumber/m2, 4) cage control - sea cucumbers and fish allowed access with as much caging material present as possible, and 5) uncaged control - sea cucumbers and fish allowed access with no caging material present. Tests showed no significant cage effects. Infauna were sampled at time 0, 1 year and 2 years, and sorted to major taxonomic group. Predator densities and feeding rates were monitored. The feeding rates of P. parvimensis varied throughout the year - the greatest amount of sediment being ingested during spring, with a significant decline during late summer through winter. Sea cucumber densities in shallow water (less than 13m) declined significantly during late summer and fall due to a downslope migration. The labrid fish had consistent feeding rates, but their densities also varied throughout the year. There was no detectable effect of holothurian deposit feeding on any of the infaunal groups. Neither normal, nor ten times normal, sea cucumber densities resulted in a significant change in either the absolute or relative abundances of the infaunal species groups. Release from labrid fish predation, however, did significantly increase the densities of the polychaete, Chaetopterus variopedatus, and the phoronid, Phoronopsis californica.

There have been increased developments in ocean exploration using autonomous underwater vehicles (AUVs) and unmanned underwater vehicles (UUVs). However, the contrast of underwater images is still a major issue for application. It is difficult to acquire clear underwater images around underwater vehicles. Since the 1960s, sonar sensors have been extensively used to detect and recognize objects in oceans. Due to the principles of acoustic imaging, sonar-imaged images have many shortcomings, such as a low signal to noise ratio and a low resolution. Consequently, vision sensors must be used for short-range identification because sonars yield to low-quality images. This thesis will concentrate solely on the optical imaging sensors for ocean observing. Although the underwater optical imaging technology makes a great progress, the recognition of underwater objects also remains a major issue in recent days. Different from the common images, underwater images suffer from poor visibility due to the medium scattering and light distortion. First of all, capturing images underwater are difficult, mostly due to attenuation caused by light. The random attenuation of the light mainly causes the haze appearance along with the part of the light scattered back from the water. In particular, the objects at a distance of more than 10 m are almost indistinguishable because of absorption. Furthermore, when the artificial light is employed, it can cause a distinctive footprint on the seafloor. In this paper, we will analysis the recent trends of ocean exploration approaches.

Repeatable and accurate seagrass mapping is required for understanding seagrass ecology and supporting management decisions. For shallow (<5 m) seagrass habitats, these maps can be created by integrating high spatial resolution imagery with field survey data. Field survey data for seagrass are often collected via snorkelling or diving. However, these methods are limited by environmental and safety considerations. Autonomous underwater vehicles (AUVs) are used increasingly to collect field data for habitat mapping, albeit mostly in deeper waters (>20 m). Here, we demonstrate and evaluate the use and potential advantages of AUV field data collection for calibration and validation of seagrass habitat mapping of shallow waters (<5 m), from multispectral satellite imagery. The study was conducted in the seagrass habitats of the Eastern Banks (142 km2), Moreton Bay, Australia. In the field, georeferenced photographs of the seagrass were collected along transects via snorkelling or an AUV. Photographs from both collection methods were analysed manually for seagrass species composition and then used as calibration and validation data to map seagrass using an established semi-automated object-based mapping routine. A comparison of the relative advantages and disadvantages of AUV and snorkeller-collected field data-sets and their influence on the mapping routine was conducted. AUV data collection was more consistent, repeatable and safer in comparison with snorkeller transects. Inclusion of deeper water AUV data resulted in mapping of a larger extent of seagrass (~7 km2, 5% of study area) in the deeper waters of the site. Although overall map accuracies did not differ considerably, inclusion of the AUV data from deeper water transects corrected errors in seagrass mapped at depths to 5 m, but where the bottom is visible on satellite imagery. Our results demonstrate that further development of AUV technology is justified for the monitoring of seagrass habitats in ongoing management programmes.

Abstract: In recent years, a potential interest has been developed in exploring and enhancing the quality of underwater images. The captured underwater images often experience scattering and high levels of noise due to the properties of light transit within the water and the biotic movement that occurs inside the ocean bottom. Underwater cameras or devices were fitted inside the unmanned underwater vehicles (UUVs), autonomous underwater vehicles (AUVs), remotely operated vehicles (ROVs), ocean detector networks, etc to see the ocean bottom. Despite being a crucial detector for keeping an eye on underwater landscapes, modern underwater camera sensors or detectors contain a number of problems. Numerous solutions to the issues with conventional underwater photography have been put out during the last five years. With the help of the number of contributions and difficulties reported by more than 40 researchers in their research article, this study focused to mention the current underwater image processing methods. This article represents an overview of numerous underwater image-processing techniques utilized, including underwater picture de-scattering, underwater image colour refurbishment, and underwater image aspect evaluations. Also, the latest trends, problems, and challenges for creating and processing underwater image sensors were discussed.

In this paper, an autonomous real-time mine detection algorithm based on improved YOLOv5 in a complex underwater environment is proposed. The target recognition system based on the optical image is used to guide autonomous underwater vehicles (AUV) to achieve high precision mine exploration and assist the autonomous mine hunting weapon system to finish the mine-hunting task. In view of all kinds of various interference information of the underwater optical images, a set of image preprocessing systems has been added to the algorithm based on YOLOv5 which has the advantages of both detection accuracy and detection speed. The original data are processed by data enhancement, image enhancement, sample equalization, negative sample addition, and Poisson fusion, which can reduce the problem of poor detection results caused by the input image quality to some extent. At the same time, the number of layers of YOLOv5 is reduced, which makes the detection algorithm meet the requirements of the AUV embedded system without affecting the detection accuracy, and improves the real-time detection of the mobile terminals. The experimental results show that the proposed method could effectively improve the detection accuracy and real-time performance of the underwater mine detection system.

Underwater Vehicles have become more sophisticated, driven by the off-shore sector and the scientific community’s rapid advancements in underwater operations. Notably, many underwater tasks, including the assessment of subsea infrastructure, are performed with the assistance of Autonomous Underwater Vehicles (AUVs). There have been recent breakthroughs in Artificial Intelligence (AI) and, notably, Deep Learning (DL) models and applications, which have widespread usage in a variety of fields, including aerial unmanned vehicles, autonomous car navigation, and other applications. However, they are not as prevalent in underwater applications due to the difficulty of obtaining underwater datasets for a specific application. In this sense, the current study utilises recent advancements in the area of DL to construct a bespoke dataset generated from photographs of items captured in a laboratory environment. Generative Adversarial Networks (GANs) were utilised to translate the laboratory object dataset into the underwater domain by combining the collected images with photographs containing the underwater environment. The findings demonstrated the feasibility of creating such a dataset, since the resulting images closely resembled the real underwater environment when compared with real-world underwater ship hull images. Therefore, the artificial datasets of the underwater environment can overcome the difficulties arising from the limited access to real-world underwater images and are used to enhance underwater operations through underwater object image classification and detection.

On Kanda port construction in Fukuoka prefecture, no fewer than 2,800 harmful chemical bombs have been discovered beneath the sea bottom so far. In advance of dredging in the port, we have carefully carried out magnetic inspection for looking for chemical bombs. Since a true target often gets mixed in a lot of suspicious metallic objects, we must cautiously draw up detected objects one by one. Use of a magnetic sensor only is uneconomical and inefficient. On account of these, we started a new three-year project of developing a sub-bottom synthetic aperture imaging sonar (sub-bottom SAS) system by using an autonomous underwater vehicle (AUV) and an autonomous surface vehicle for tracking the AUV. The AUV Tri-Dog 1, which was developed by T. Ura and H. Kondo, is going to be used. The AUV will be installed with a new sub-bottom SAS system and could perform stable and suitable navigation for SAS imaging in shallow water. The surface autonomous tracking vehicle can compensate for positioning error of the AUV and monitor sub-bottom imaging. The synthetic aperture sonar technology for sub-bottom imaging is now in the worldwide limelight, and several projects are ongoing. Since most of chemical bombs were buried up to 5 meters beneath the sea bottom, operating frequency 10 kHz to 30 kHz is expected to be useful for sub-bottom SAS. At the beginning of the project, we measured acoustical characteristics of sub-bottom sediments in Kanda port in 2010 May using a 33kHz single beam echo sounder RESON-Navisound 410 and a parametric sub-bottom profiler Innomar SES2000 Compact. Based on the acoustical characteristics, we will design a sonar projector and hydrophone arrays and a new surface autonomous tracking vehicle in this year.

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper OTC 35821, “A Mobile Platform for AI-Powered Underwater Image Processing,” by Chenyu Zhou, University of Houston; Lily Walker, William &amp; Mary; and Xin Fu, University of Houston, et al. The paper has not been peer reviewed. Copyright 2025 Offshore Technology Conference. _ Timely inspection of subsea infrastructure—particularly pipelines—is crucial for preventing oil spills and protecting the environment. This paper introduces a technology for offshore pipeline inspection centered on an autonomous robotic system equipped with underwater computer vision and edge-computing capabilities. The aim is to create a fast, cost-effective solution for inspecting underwater structures, enabling early detection of environmental hazards. Introduction Autonomous underwater vehicles (AUVs) have revolutionized the ability to explore and monitor marine environments. The cornerstone of AUV capabilities lies in their imaging systems, although the underwater environment poses unique and complex challenges for these systems. Additional image-processing and correction ability remains a goal for AUVs, creating computational demands that must be balanced against limited onboard power and processing capabilities (Fig. 1). Recently, advancements in deep learning have enabled deep neural networks (DNNs) to achieve remarkable success in solving a wide range of image-processing challenges, significantly surpassing the performance of traditional methods. In this research, the authors build upon foundational work in the literature to present a comprehensive workflow for adapting and deploying pretrained DNN models onto mobile computing platforms suitable for integration with AUVs. The described approach examines and documents the transition from powerful desktop systems to resource-constrained mobile platforms. Underwater Image Enhancement Motion blur is a common challenge in underwater image processing, especially when the camera is mounted on a moving platform such as an AUV. The lack of adequate lighting in underwater environments often requires longer exposure times, further exacerbating motion blur. The blurring process can be mathematically described as a convolution of the sharp image with a blur kernel, plus noise. If the blur kernel is known beforehand, the problem becomes one of nonblind deconvolution. However, when the blur kernel is unknown—common in dynamic underwater settings—it becomes a blind deconvolution problem. This problem is inherently ill-posed because multiple combinations of sharp images and blur kernels can explain the same blurred image. To resolve this ambiguity, assumptions or previous knowledge about the blur kernel or image structure typically are introduced. Traditional blind deconvolution methods often use iterative algorithms, refining the blur kernel and sharp image estimates over multiple steps. These classical approaches, however, often suffer from artifacts such as ringing and struggling with large motions or noise. To address these limitations, researchers have explored regularization techniques that incorporate priors to constrain the solution space. For example, sparsity priors in the temporal domain encourage smooth results, while Bayesian frameworks provide probabilistic formulations of the problem. One promising direction involves leveraging domain-specific priors for underwater imagery. By incorporating this property into optimization frameworks, researchers have developed more-effective and -robust blind deblurring methods tailored to underwater scenarios.

An autonomous underwater vehicle (AUV) needs to be fitted with a forward looking sonar designed to work in the difficult acoustic conditions encountered in shallow water environments. Detection and classification of objects in shallow water requires the sonar to provide significantly more acoustic resolution than is available with current in-service small vehicle sonars. Consequently, design of a shallow water sonar will involve marked departures from prior systems in terms of operating frequency, array geometry and signal processing requirements. A notable consequence is that the acoustic array of a high resolution imaging sonar will contain hundreds of sensors. The feasibility of implementing hundreds of sensors in an AUV sonar will depend on finding a sensor that meets requisite technical requirements and is affordable. This paper describes and discusses two candidate acoustic sensor technologies, one of which has already been incorporated into a high resolution array and one that has been tested in a prototype array and is currently being fabricated into a full size array.

A potential feeding regulation strategy during aestivation: Relaxation of intestine mediated by pedal peptide/orcokinin-type neuropeptides in sea cucumber Apostichopus japonicus

Sea Cucumbers Linette Marie Allen (bio) The sea where living creatures were at one time immersed is now enclosed within their bodies. —Italo Calvino Surely Calvino's right. If a sea cucumber doesn't enclose the sea within her body, then I can't name a creature who does. This crawler I've just lifted out of shallow water, in fact, is blob-like and leathery. Her complexion is a color bomb—clefts of clementine nailing nooks of nonda plum. When I turn her toward the moon, rays of pearl pellicle her backside. In my wet hands she feels roughly five pounds and looks the length of a rainbow lobster. She's been vacuuming oceans for years, traces of sea-mint inside, and poop, which she excretes generously onto sea floors, bowels like sage & tomato sausages cold-pressed in Italian kitchens. I tickle her tentacles and watch her withdraw. If I were Dalí, I'd paint her as a warped pickle on a Paris clothesline, her imaginary bloomers wetting wildflowers: corn-cockle, dog-violet, fireweed. Though the pearlfish and the Harlequin crab make their home in her rectum, I'd choose not to paint them. I might, however, feature her cousins—the sand dollar and the starfish—for the sea, it seems to me, must have everything in the world to do with stars, to do with fish, trickling sands. Sea cucumbers are omnivores. They eat with their mouths and their anuses. Sensing food on the surface of the seabed, a sea cucumber will burst forth its tentacles and yank microscopic morsels toward its opened head. Although sea cucumbers are brainless, heartless, bloodless, boneless, and blind, they are gluttons. I try to imagine a juvenile spaghetti worm, wonder whether it can feel the nerve ring spreading, sense the presence of something hidden lurking, hear anything at all when fingers lunge like starved children, about to offer a sacrifice. A juvenile spaghetti worm is translucent and lives among ocean drifters near the seabed. Smaller than a grain of rice, and probably polluted with tiny shreds of plastic in its backside, it has only the ability to pray away the danger, to wish it were a scorpion, an angler, a monster. And then the sea cucumber's hand—like the lacy treetops in a prized Paris watercolor—snatches it and sucks the life out of it. Its long arm bends to its mouth and releases what's left of the structure. The juice satisfies. It makes the cucumber feel alive again. But it doesn't stop there. The hand is accompanied by nine other hands and they go ballistic, gobbling up the worm's brothers and sisters and cousins—all young; and fiery; and mighty-fine looking. [End Page 140] What does it sound like, that initial suck? Does skin swoosh and swizzle like a toilet flushing after supper, like gale winds rustling bluebells? A flushing in slow motion like bricks breaking windows or belts beating lovebirds or needles violating vinyl, vinyl hissing and weeping for secretaries who'd always wanted to be singers—a fat kick to big knees around midnight? Perhaps my painting must be an illusion. With fresh sheet of canvas, white and clean and expecting, the scene re-cast: marble café table lampside with bowl of soup. Yes. Sea cucumber as lamp, her bowels as light, ocean drifters as marble, coral reef as table, and—because even an illusion must embody realism—pearlfish as soup. To capture the threat of the sea, I could certainly throw in a sea turtle or a jellyfish. Sea cucumbers are terrified of this lot, especially the latter, whether the Flower Hats or Cabbage Heads or Cannonballs. Any predator will trigger a 'fight or flight' response in the cucumber. It will flee into the belly of the coral, holding its breath, until the danger passes. If startled, it'll shoot its intestines through its anus. This shooting is white, stringy, sticky, and toxic. But it can't hold its breath forever and the pearlfish knows it. It knows that the cucumber's rectum is multi-purpose: for breathing, for reproducing, for absorbing nutrients from food. And if the pearlfish can't have its...

Introduction Effective underwater vision is critical for real-time marine ecosystem observation and conservation, especially for autonomous underwater vehicles (AUVs) operating in challenging oceanic environments. Methods We propose a novel underwater image enhancement framework tailored for smart robotic systems used in biodiversity monitoring, habitat mapping, and environmental sensing. Our method integrates a Denoising Diffusion Probabilistic Model (DDPM) for progressive image restoration with an Attention-Enhanced Convolutional Blocks (AECB) augmented Transformer backbone. The AECB modules provide dual channel and spatial attention, selectively amplifying features to enhance visual quality. Additionally, a lightweight architecture combined with a skip-sampling strategy is designed to optimize computational efficiency for onboard deployment in AUVs and underwater drones. Results Experimental evaluations demonstrate that our framework achieves superior image restoration performance while maintaining computational efficiency, outperforming existing transformer-diffusion approaches. The dual attention mechanism within AECB modules distinctly improves the clarity and detail of underwater images. Discussion This work advances AI-driven perception systems for intelligent ocean observation technologies, supporting improved marine biodiversity protection. The proposed model promises practical real-time applications in autonomous underwater exploration and monitoring. The model and code will be made publicly available on GitHub: https://github.com/ntiwari91/DM-AECB .

Underwater cables or pipelines are commonly utilized elements in ocean research, marine engineering, power transmission, and communication-based activities. Their performance necessitates regularly conducted inspection for maintenance purposes. A vision system is commonly used by autonomous underwater vehicles (AUVs) to track and search for underwater cable. Its traditional methods are characteristically applicable in AUVs, wherein they are equipped with handcrafted features and shallow trainable architectures. However, such methods are subpar or even incapable of tracking underwater cable in fast-changing and complex underwater conditions. In contrast to this, the deep learning method is linked with the capacity to learn semantic, high-level, and deeper features, thus rendering it recommended for performing underwater cable tracking. In this study, several deep Convolutional Neural Network (CNN) models were proposed to classify underwater cable images obtained from a set of underwater images, whereby transfer learning and data augmentation were applied to enhance the classification accuracy. Following a comparison and discussion regarding the performance of these models, MobileNetV2 outperformed among other models and yielded lower computational time and the highest accuracy for classifying underwater cable images at 93.5%. Hence, the main contribution of this study is geared toward developing a deep learning method for underwater cable image classification.