Return to search results
Automated Fish Detection and Wildlife Tracking for Marine Energy Applications: Final Report and Review
This submission contains two technical resources produced for the U.S. Department of Energy Water Power Technologies Office (WPTO) that evaluate and advance automated methods for underwater wildlife detection, tracking, and monitoring around marine energy systems. The project analyzes how modern computer-vision and machine-learning techniques can support reductions in manual labor, improve monitoring consistency, and streamline permitting by enabling automated detection of fish in both optical and imaging-sonar data.
The work includes development and testing of You Only Look Once (YOLO), Faster Region-based Convolutional Neural Network (R-CNN), Elastic Shape Analysis (ESA), and hyper-image methods for fish detection using the Pacific Northwest National Laboratory (PNNL) EyeSea optical dataset and multiple imaging sonar datasets. It also provides a comprehensive literature review on imaging sonar performance, machine-learning approaches for object detection, and domain-adaptation strategies to support cross-site generalization.
Resources
2 resources available
-
Fish Detection AI final report.docx
DOCX -
Automatic Underwater Wildlife Tracking with Imaging Sonar Review.docx
DOCX
Find Related Datasets
Search by Tags
Click any tag below to search for similar datasets
Complete Metadata
| @type | dcat:Dataset |
|---|---|
| accessLevel | public |
| bureauCode |
[ "019:20" ] |
| contactPoint |
{ "fn": "Denis-Marc Nault", "@type": "vcard:Contact", "hasEmail": "mailto:denis-marc.nault@doe.gov" } |
| dataQuality |
true
|
| description | This submission contains two technical resources produced for the U.S. Department of Energy Water Power Technologies Office (WPTO) that evaluate and advance automated methods for underwater wildlife detection, tracking, and monitoring around marine energy systems. The project analyzes how modern computer-vision and machine-learning techniques can support reductions in manual labor, improve monitoring consistency, and streamline permitting by enabling automated detection of fish in both optical and imaging-sonar data. The work includes development and testing of You Only Look Once (YOLO), Faster Region-based Convolutional Neural Network (R-CNN), Elastic Shape Analysis (ESA), and hyper-image methods for fish detection using the Pacific Northwest National Laboratory (PNNL) EyeSea optical dataset and multiple imaging sonar datasets. It also provides a comprehensive literature review on imaging sonar performance, machine-learning approaches for object detection, and domain-adaptation strategies to support cross-site generalization. |
| distribution |
[ { "@type": "dcat:Distribution", "title": "Fish Detection AI final report.docx", "format": "docx", "accessURL": "https://mhkdr.openei.org/files/675/AL_Fish%20Detection%20AI%20final%20report_V3%20%281%29.docx", "mediaType": "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "description": "The final technical report detailing development, evaluation, and comparison of multiple AI/computer-vision methods (YOLO, Faster R-CNN, ESA, hyper-images) for automated fish detection and counting using optical and imaging-sonar data. Includes model inputs, preprocessing methods, experiments, results, and recommendations." }, { "@type": "dcat:Distribution", "title": "Automatic Underwater Wildlife Tracking with Imaging Sonar Review.docx", "format": "docx", "accessURL": "https://mhkdr.openei.org/files/675/Automatic%20Underwater%20Wildlife%20Tracking%20with%20Imaging%20Sonar_whitepaper_WPTOReview%20%281%29.docx", "mediaType": "application/vnd.openxmlformats-officedocument.wordprocessingml.document", "description": "The literature review and recommendations describing imaging sonar systems, wildlife monitoring applications, AI/ML techniques for sonar image analysis, and domain-adaptation approaches to improve detection robustness across different sites and environmental conditions." } ] |
| identifier | https://data.openei.org/submissions/8753 |
| issued | 2025-12-08T07:00:00Z |
| keyword |
[ "ESA", "Hydrokinetic", "MHK", "Marine", "PNNL EyeSea", "R-CNN", "YOLO", "You Only Look Once", "automated fish detection", "computer-vision", "elasric shape analysis", "energy", "hyper-image", "imaging sonar", "literature review", "machine-learning", "object tracking", "optical dataset", "power", "wildlife tracking" ] |
| landingPage | https://mhkdr.openei.org/submissions/675 |
| license | https://creativecommons.org/licenses/by/4.0/ |
| modified | 2026-08-17T16:28:44Z |
| programCode |
[ "019:009" ] |
| publisher |
{ "name": "Booz Allen Hamilton Inc. ", "@type": "org:Organization" } |
| spatial |
"{"type":"Polygon","coordinates":[[[-180,-83],[180,-83],[180,83],[-180,83],[-180,-83]]]}"
|
| title | Automated Fish Detection and Wildlife Tracking for Marine Energy Applications: Final Report and Review |
