Mapping and quantifying pelagic Sargassum in the Atlantic Ocean using multi-band medium-resolution satellite data and deep learning

Remote Sensing of Environment, Год журнала: 2023, Номер 289, С. 113515 - 113515

Опубликована: Фев. 28, 2023

Язык: Английский

---

Assessing the efficiency of pixel-based and object-based image classification using deep learning in an agricultural Mediterranean plain

Environmental Monitoring and Assessment, Год журнала: 2025, Номер 197(2)

Опубликована: Янв. 10, 2025

Язык: Английский

---

Potential of seagrass habitat restorations as nature-based solutions: Practical and scientific implications in Indonesia

AMBIO, Год журнала: 2022, Номер 52(3), С. 546 - 555

Опубликована: Дек. 9, 2022

Seagrasses offer diverse ecosystem services, yet, they are among the most threatened ecosystems. When degraded or destroyed, their services lost reduced in process, affecting, for instance, local communities directly dependent on livelihood provision. The Intergovernmental Panel Climate Change (IPCC) reported that climate change is projected to worsen over time; thus, there an urgent need mitigation strategies practice and also longer term. This work aims provide alternative perspective of seagrass restoration as a nature based solution (NbS) global scale, giving emphasis tropical regions such Indonesia. We focused restorations which not yet well established comparison with other programs (e.g., mangroves) despite critical roles. present this how restoring meadows fits standard NbS published by International Union Conservation Nature (IUCN). results study can serve basis promoting against particularly countries wide extent coverage.

Язык: Английский

---

The Dynamics of Seagrass Ecosystems: History, Past Accomplishments, and Future Prospects

Estuaries and Coasts, Год журнала: 2023, Номер 46(7), С. 1653 - 1676

Опубликована: Сен. 5, 2023

Abstract The goal of this perspective is to review how seagrass research has evolved over the past half century our current state knowledge. We knowledge ecosystems changed from pre-1970s when so little was known about seagrasses and it progressed during next 5 decades rapidly expanded. Here, we concentrate on accomplishments in areas reproductive biology ecology, population biology, seagrass-animal relationships, conservation restoration, mapping monitoring. also look ahead discuss some that are ripe for future research, especially those employing new monitoring technologies, improved restoration methods include multiple genetic variants, rhizosphere studies result a deeper understanding microbial effects nitrogen availability, sulfide levels carbon sequestration, changing climatic regimes tropicalization will likely affect temperate tropical seagrass-dominated ecosystems.

Язык: Английский

---

Image Labels Are All You Need for Coarse Seagrass Segmentation

2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Год журнала: 2024, Номер unknown, С. 5931 - 5940

Опубликована: Янв. 3, 2024

Seagrass meadows serve as critical carbon sinks, but estimating the amount of they store requires knowledge seagrass species present. Underwater and surface vehicles equipped with machine learning algorithms can help to accurately estimate composition extent at scale. However, previous approaches for detection classification have required supervision from patch-level labels. In this paper, we reframe a weakly supervised coarse segmentation problem where image-level labels are used during training (25 times fewer compared labeling) outputs obtained inference time. To end, introduce SeaFeats, an architecture that uses unsupervised contrastive pre-training feature similarity, SeaCLIP, model showcases effectiveness large language models supervisory signal in domain-specific applications. We demonstrate ensemble SeaFeats SeaCLIP leads highly robust performance. Our method outperforms require on multi-species 'DeepSeagrass' dataset by 6.8% (absolute) class-weighted F1 score, 12.1% presence/absence score 'Global Wetlands' dataset. also present two case studies real-world deployment: outlier Global Wetlands dataset, application our imagery collected FloatyBoat autonomous vehicle.

Язык: Английский

---

Capturing the Dynamics of Aboveground Carbon Stock in Intertidal Seagrass Meadows using Sentinel-2 Time-Series Imagery

Remote Sensing Applications Society and Environment, Год журнала: 2025, Номер unknown, С. 101552 - 101552

Опубликована: Апрель 1, 2025

Язык: Английский

---

Rapid loss of temperate kelp forests revealed by unmanned aerial vehicle (UAV) photography and underwater observations

Aquatic Botany, Год журнала: 2025, Номер 200, С. 103900 - 103900

Опубликована: Апрель 29, 2025

Язык: Английский

---

A Survey of Deep Learning Approaches for the Monitoring and Classification of Seagrass

Ocean Science Journal, Год журнала: 2025, Номер 60(2)

Опубликована: Май 15, 2025

Язык: Английский

---

Biomass estimations of cultivated kelp using underwater RGB images from a mini-ROV and computer vision approaches

Frontiers in Marine Science, Год журнала: 2024, Номер 11

Опубликована: Март 12, 2024

Seaweed farming is the fastest-growing aquaculture sector worldwide. As farms continue to expand, automated methods for monitoring growth and biomass become increasingly important. Imaging techniques, such as Computer Vision (CV), which allow automatic object detection segmentation can be used rapid estimation of underwater kelp size. Here, we segmented in situ RGB images cultivated Saccharina latissima using CV techniques explored pixel area a tool estimations. Sampling consisted imaging S. hanging vertically from cultivation line mini-ROV. In chlorophyll concentrations turbidity (proxies phytoplankton particle concentrations) were monitored water visibility. We first compared manual length estimations individuals obtained (through annotation ImageJ software). Then, applied segment calculate investigated these measurements robust proxy wet weight biomass. A strong positive linear correlation (r 2 = 0.959) between estimates image frames harvested was observed. Using unsupervised learning algorithms, mean shift clustering, colour adaptive thresholding OpenCV package Python, number individual pixels contour counted. power relationship found with CV-derived 0.808) estimated images. Likewise, had (r² 0.887). When removing data where visibility poor due high levels (mid-June), stronger field 0.976 r² 0.979 length). These results show that are possible through segmentation. However, demonstrate quality post-processing accuracy model highly dependent environmental conditions (e.g. concentrations). The establishment technologies has potential offer scalability production, efficient real-time sea improved yield predictions.

Язык: Английский

---

High-Resolution Seagrass Species Mapping and Propeller Scars Detection in Tanjung Benoa, Bali through UAV Imagery

Journal of Ecological Engineering, Год журнала: 2023, Номер 25(1), С. 161 - 174

Опубликована: Дек. 2, 2023

As a part of the marine ecosystem, seagrass plays significant role in coastal environment.However, due to increased threats from natural causes and anthropogenic pressures, decline will likely begin many areas world.Therefore, several studies have been carried out observe distribution help resolve issue.Remote sensing is often used its ability achieve high accuracy when distinguishing distribution.Still, this method lacks species classification because not all satellites similar aerial vehicles fine spatial resolution distinguish distinct seagrass.In study, we aim address issue by utilizing unmanned (UAV), which are known for providing finer better imagery.Samuh Beach at Tanjung Benoa, Bali, Indonesia, was chosen as study site location it experiences levels tourism activities.From UAV flight mission, images obtained were processed.The result's also tested with an error matrix.The found Enhalus acoroides, Halodule pinifolia, Thalassia hemprichii, Cymodocea rotundata, Syringodium isoetifolium, 65% overall map.This result indicates that UAVs can be strong option future.In addition that, able scars on beds left boat propeller activities tourism.However, further research needed gain understanding these objects.

Язык: Английский

---