Surveying the deep: A review of computer vision in the benthos

Ecological Informatics, Journal Year: 2025, Volume and Issue: unknown, P. 102989 - 102989

Published: Jan. 1, 2025

Language: Английский

---

Exposing inequities in deep-sea exploration and research: results of the 2022 Global Deep-Sea Capacity Assessment

Frontiers in Marine Science, Journal Year: 2023, Volume and Issue: 10

Published: Aug. 9, 2023

The 2022 Global Deep-Sea Capacity Assessment is a baseline assessment of the technical and human capacity for deep-sea exploration research in every coastal area with deep ocean worldwide. From 200 to nearly 11,000 meters below sea level, encompasses single largest—and arguably most critical—biosphere on Earth. Globally, two-thirds all exclusive economic zones combined have water depths between 2,000 6,000 meters, making this particularly critical depth range access. This study includes information 186 countries territories, analyzed by subregional, regional, income groups. data were collected through both an online survey manual research. We found that globally, 52% respondents agreed considered important their community. A third they had in-country technology conduct research, half expertise. Survey results revealed challenges worldwide are funding, access vessels, capacity. top three global opportunities training opportunities, less expensive collection technology, better analysis tools. provides necessary strategically develop, equitably implement, quantitatively measure impact development over coming years. It now possible evolution next decade, indicator progress during UN Decade Ocean Science Sustainable Development.

Language: Английский

---

Systematic Distribution of Bioluminescence in Marine Animals: A Species-Level Inventory

Life, Journal Year: 2024, Volume and Issue: 14(4), P. 432 - 432

Published: March 24, 2024

Bioluminescence is the production of visible light by an organism. This phenomenon particularly widespread in marine animals, especially deep sea. While luminescent status numerous animals has been recently clarified thanks to advancements deep-sea exploration technologies and phylogenetics, that others become more obscure due dramatic changes systematics (themselves triggered molecular phylogenies). Here, we combined a comprehensive literature review with unpublished data establish catalogue animals. Inventoried were identified species level over 97% cases associated score reflecting robustness their luminescence record. capability established 695 genera reports from 99 additional need further confirmation. Altogether, these potentially encompass 9405 species, which 2781 are luminescent, 136 (e.g., suggested those needs confirmation), non-luminescent, 6389 have unknown status. Comparative analyses reveal new insights into occurrence among animal groups highlight promising research areas. work will provide solid foundation for future studies related field bioluminescence.

Language: Английский

---

Artificial intelligence for mineral exploration: A review and perspectives on future directions from data science

Earth-Science Reviews, Journal Year: 2024, Volume and Issue: unknown, P. 104941 - 104941

Published: Sept. 1, 2024

Language: Английский

---

Towards equity and justice in ocean sciences

npj Ocean Sustainability, Journal Year: 2023, Volume and Issue: 2(1)

Published: Dec. 13, 2023

Abstract The global scientific community is currently going through a self-reckoning in which it questioning and re-examining its existing practices, many of are based on colonial neo-colonial perceptions. This particularly acute for the ocean research community, where unequal unbalanced international collaborations have been rife. Consequently, numerous discussions calls made to change current status quo by developing guidelines frameworks addressing key issues plaguing our community. Here, we provide an overview topics that has debated over last three four years, with emphasis research, coupled actions per stakeholder groups (research institutions, funding agencies, publishers). We also outline some missing suggest path forward tackle these gaps. hope this contribution will further accelerate efforts bring more equity justice into sciences.

Language: Английский

---

Nu—A Marine Life Monitoring and Exploration Submarine System

Technologies, Journal Year: 2025, Volume and Issue: 13(1), P. 41 - 41

Published: Jan. 20, 2025

Marine life exploration is constrained by factors such as limited scuba diving time, depth restrictions for divers, costly expeditions, safety risks to divers’ health, and minimizing harm marine ecosystems, where traditional often disturbing life. This paper introduces Nu (named after an ancient Egyptian deity), a 3D-printed Remotely Operated Underwater Vehicle (ROUV) designed in attempt address these challenges. employs Long Range (LoRa), low-power long-range communication technology, enabling wireless operation via manual controller. The vehicle features onboard live-feed camera with separate system that transmits video external real-time machine learning (ML) pipeline fish species classification, reducing human error taxonomists. It uses Brushless Direct Current (BLDC) motors long-distance movement water pump precise navigation, disturbance, damage surrounding species. Nu’s functionality was evaluated controlled 2.5-m-deep body of water, focusing on connectivity, maneuverability, identification accuracy. detection algorithm achieved average precision 60% identifying presence, while the classification model 97% assigning labels, unknown flagged correctly. testing environment has met design expectations.

Language: Английский

---

Geoportals in marine spatial planning: state of the art and future perspectives

Ocean & Coastal Management, Journal Year: 2025, Volume and Issue: 266, P. 107688 - 107688

Published: April 15, 2025

Language: Английский

---

Using open-access mapping interfaces to advance deep ocean understanding

CSI Transactions on ICT, Journal Year: 2025, Volume and Issue: unknown

Published: March 15, 2025

Language: Английский

---

Iterative mapping of marine ecosystems for spatial status assessment, prioritization, and decision support

Frontiers in Ecology and Evolution, Journal Year: 2023, Volume and Issue: 11

Published: March 31, 2023

South Africa has taken an iterative approach to marine ecosystem mapping over 18 years that provided a valuable foundation for assessment, planning and decision-making, supporting improved ecosystem-based management protection. Iterative progress been made in overcoming challenges faced by developing countries, especially the inaccessible realm. Our aim is report on produce improve national map guide other countries facing similar challenges, illustrate impact of even simplest map. produced four versions, from rudimentary 34 biozones informed bathymetry data, latest version comprising 163 types 83 environmental biodiversity datasets aligns with IUCN Global Ecosystem Typology. Data were unlocked through academic industry collaborations; multi-disciplinary, multi-realm multi-generational networks practitioners; targeted research address key gaps. To advance toward more transparent, reproducible data-driven approach, limitations, barriers opportunities improvement identified. Challenges included limited human data infrastructure capacity collate, curate assimilate many sources, covering variety components, methods scales. Five lessons are relevance others working classification mapping, distilled. These include (1) benefits improvement; (2) value fostering relationships among co-ordinated network practitioners including early-career researchers; (3) strategically prioritizing leveraging resources build foundational understand drivers pattern; (4) need developing, transferring applying tools enhance quality, analytical workflows outputs; (5) application new technology emerging statistical prediction pattern. Africa’s successfully applied spatial prioritization support protected area expansion planning. successes demonstrate who continually evidence base iteratively while simultaneously growing ecological knowledge informing changing priorities policy.

Language: Английский

---

Deepdive: Leveraging Pre-trained Deep Learning for Deep-Sea ROV Biota Identification in the Great Barrier Reef

Scientific Data, Journal Year: 2024, Volume and Issue: 11(1)

Published: Sept. 3, 2024

Understanding and preserving the deep sea ecosystems is paramount for marine conservation efforts. Automated object (deep-sea biota) classification can enable creation of detailed habitat maps that not only aid in biodiversity assessments but also provide essential data to evaluate ecosystem health resilience. Having a significant source labelled helps prevent overfitting enables training learning models with numerous parameters. In this paper, we contribute establishment deep-sea remotely operated vehicle (ROV) image dataset 3994 images featuring biota belonging 33 classes. We manually label through rigorous quality control human-in-the-loop labelling. Leveraging from ROV equipped advanced imaging systems, our study provides results using novel deep-learning classification. use including ResNet, DenseNet, Inception, Inception-ResNet benchmark features class imbalance many Our show model mean accuracy 65%, AUC scores exceeding 0.8 each class.

Language: Английский

---