{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,11]],"date-time":"2026-07-11T19:25:14Z","timestamp":1783797914022,"version":"3.55.0"},"reference-count":33,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,10,17]],"date-time":"2025-10-17T00:00:00Z","timestamp":1760659200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Robot. AI"],"abstract":"<jats:p>Due to their utility in replacing workers in tasks unsuitable for humans, unmanned underwater vehicles (UUVs) have become increasingly common tools in the fish farming industry. However, earlier studies and anecdotal evidence from farmers imply that farmed fish tend to move away from and avoid intrusive objects such as vehicles that are deployed and operated inside net pens. Such responses could imply a discomfort associated with the intrusive objects, which, in turn, can lead to stress and impaired welfare in the fish. To prevent this, vehicles and their control systems should be designed to automatically adjust operations when they perceive that they are repelling the fish. A necessary first step in this direction is to develop on-vehicle observation systems for assessing object\/vehicle\u2013fish distances in real-time settings that can provide inputs to the control algorithms. Due to their small size and low weight, modern cameras are ideal for this purpose. Moreover, the ongoing rapid developments within deep learning methods are enabling the use of increasingly sophisticated methods for analyzing footage from cameras. To explore this potential, we developed three new pipelines for the automated assessment of fish\u2013camera distances in video and images. These methods were complemented using a recently published method, yielding four pipelines in total, namely, <jats:italic>SegmentDepth<\/jats:italic>, <jats:italic>BBoxDepth<\/jats:italic>, and <jats:italic>SuperGlue<\/jats:italic> that were based on stereo-vision and <jats:italic>DepthAnything<\/jats:italic> that was monocular. The overall performance was investigated using field data by comparing the fish\u2013object distances obtained from the methods with those measured using a sonar. The four methods were then benchmarked by comparing the number of objects detected and the quality and overall accuracy of the stereo matches (only stereo-based methods). <jats:italic>SegmentDepth<\/jats:italic>, <jats:italic>DepthAnything<\/jats:italic>, and <jats:italic>SuperGlue<\/jats:italic> performed well in comparison with the sonar data, yielding mean absolute errors (MAE) of 0.205 m (95% CI: 0.050\u20130.360), 0.412 m (95% CI: 0.148\u20130.676), and 0.187 m (95% CI: 0.073\u20130.300), respectively, and were integrated into the Robot Operating System (ROS2) framework to enable real-time application in fish behavior identification and the control of robotic vehicles such as UUVs.<\/jats:p>","DOI":"10.3389\/frobt.2025.1628213","type":"journal-article","created":{"date-parts":[[2025,10,17]],"date-time":"2025-10-17T07:41:07Z","timestamp":1760686867000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":4,"title":["Deep learning methods for 3D tracking of fish in challenging underwater conditions for future perception in autonomous underwater vehicles"],"prefix":"10.3389","volume":"12","author":[{"given":"Martin","family":"F\u00f8re","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emilia May","family":"O\u2019Brien","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eleni","family":"Kelasidi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,10,17]]},"reference":[{"key":"B1","article-title":"A novel computer vision approach for assessing fish responses to intrusive objects in aquaculture","author":"Alvheim","year":"2025"},{"key":"B2","doi-asserted-by":"publisher","first-page":"70","DOI":"10.1109\/TFR.2024.3421393","article-title":"Three-dimensional obstacle avoidance and path planning for unmanned underwater vehicles using elastic bands","volume":"1","author":"Amundsen","year":"2024","journal-title":"IEEE Trans. Field Robotics"},{"key":"B3","doi-asserted-by":"publisher","first-page":"631","DOI":"10.1080\/08927014.2019.1640214","article-title":"Biofouling in marine aquaculture: a review of recent research and developments","volume":"35","author":"Bannister","year":"2019","journal-title":"Biofouling"},{"key":"B4","doi-asserted-by":"publisher","first-page":"867","DOI":"10.1115\/OMAE2004-51577","article-title":"Design rules for marine fish farms in Norway: calculation of the structural response of such flexible structures to verify structural integrity","volume":"37459","author":"Berstad","year":"2004","journal-title":"Int. Conf. Offshore Mech. Arct. Eng."},{"key":"B5","doi-asserted-by":"crossref","first-page":"1","DOI":"10.23919\/OCEANS.2015.7404486","article-title":"Exposed aquaculture in Norway","volume-title":"OCEANS 2015-MTS\/IEEE Washington","author":"Bjelland","year":"2015"},{"key":"B6","doi-asserted-by":"publisher","first-page":"e12964","DOI":"10.1111\/raq.12964","article-title":"Exposed aquaculture operations: strategies for safety and fish welfare","volume":"17","author":"Bjelland","year":"2024","journal-title":"Rev. Aquac."},{"key":"B7","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/IST.2017.8261525","article-title":"Designing a small-sized autonomous underwater vehicle architecture for regular periodic fish-cage net inspection","volume-title":"2017 IEEE international conference on imaging systems and techniques (IST)","author":"Chalkiadakis","year":"2017"},{"key":"B8","unstructured":"Food and agriculture organization of the united nations: aquaculture production: quantities 1950-2022\n          \n          \n          2024"},{"key":"B9","doi-asserted-by":"publisher","first-page":"176","DOI":"10.1016\/j.biosystemseng.2017.10.014","article-title":"Precision fish farming: a new framework to improve production in aquaculture","volume":"173","author":"F\u00f8re","year":"2018","journal-title":"Biosyst. Eng."},{"key":"B10","doi-asserted-by":"publisher","first-page":"109609","DOI":"10.1016\/j.compag.2024.109609","article-title":"Autonomous net inspection and cleaning in sea-based fish farms: a review","volume":"227","author":"Fu","year":"2024","journal-title":"Comput. Electron. Agric."},{"key":"B11","volume-title":"Deep-learning algorithms for estimation of fish-population parameters from stereo vision video data","author":"Jakobsen","year":"2023"},{"key":"B12","article-title":"Ultralytics yolov8","author":"Jocher","year":"2023"},{"key":"B13","first-page":"1","volume-title":"Robotics for sea-based fish farming","author":"Kelasidi","year":"2022"},{"key":"B14","doi-asserted-by":"publisher","first-page":"401","DOI":"10.1016\/j.ifacol.2022.10.461","article-title":"Autonomous monitoring and inspection operations with UUVs in fish farms","volume":"55","author":"Kelasidi","year":"2022","journal-title":"IFAC-PapersOnLine"},{"key":"B15","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV51070.2023.00371","article-title":"Segment anything","author":"Kirillov","year":"2023"},{"key":"B16","doi-asserted-by":"publisher","first-page":"108306","DOI":"10.1109\/ACCESS.2023.3320949","article-title":"Fish monitoring in aquaculture using multibeam echosounders and machine learning","volume":"11","author":"Kristmundsson","year":"2023","journal-title":"IEEE Access"},{"key":"B17","doi-asserted-by":"publisher","first-page":"191220","DOI":"10.1098\/rsos.191220","article-title":"Salmon behavioural response to robots in an aquaculture sea cage","volume":"7","author":"Kruusmaa","year":"2020","journal-title":"R. Soc. Open Sci."},{"key":"B18","doi-asserted-by":"publisher","first-page":"108099","DOI":"10.1016\/j.compag.2023.108099","article-title":"Where\u2019s your head at? Detecting the orientation and position of pigs with rotated bounding boxes","volume":"212","author":"Liu","year":"2023","journal-title":"Comput. Electron. Agric."},{"key":"B19","doi-asserted-by":"publisher","first-page":"1856","DOI":"10.1098\/rsif.2012.0084","article-title":"Fish and robots swimming together: attraction towards the robot demands biomimetic locomotion","volume":"9","author":"Marras","year":"2012","journal-title":"J. R. Soc. Interface"},{"key":"B20","doi-asserted-by":"crossref","DOI":"10.3389\/fphys.2020.00162","article-title":"Welfare and stressors in fish: challenges facing aquaculture","author":"Martos-Sitcha","year":"2020"},{"key":"B21","doi-asserted-by":"publisher","first-page":"E57","DOI":"10.1111\/j.1750-3841.2007.00626.x","article-title":"Computer vision-based evaluation of pre-and postrigor changes in size and shape of Atlantic cod (Gadus morhua) and Atlantic salmon (Salmo Salar) fillets during rigor mortis and ice storage: effects of perimortem handling stress","volume":"73","author":"Misimi","year":"2008","journal-title":"J. food Sci."},{"key":"B22","doi-asserted-by":"crossref","first-page":"1002","DOI":"10.1109\/MED48518.2020.9183157","article-title":"Analysis of a novel autonomous underwater robot for biofouling prevention and inspection in fish farms","volume-title":"2020 28th mediterranean conference on control and automation (MED)","author":"Ohrem","year":"2020"},{"key":"B23","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1117\/12.3023414","article-title":"Stereoyolo+ deepsort: a framework to track fish from underwater stereo camera in situ","volume":"13072","author":"Saad","year":"2024","journal-title":"Sixt. Int. Conf. Mach. Vis. (ICMV 2023)"},{"key":"B24","doi-asserted-by":"publisher","first-page":"369","DOI":"10.1111\/raq.12143","article-title":"Application of machine vision systems in aquaculture with emphasis on fish: State-of-the-art and key issues","volume":"9","author":"Saberioon","year":"2017","journal-title":"Rev. Aquac."},{"key":"B25","first-page":"4938","article-title":"Superglue: learning feature matching with graph neural networks","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Sarlin","year":"2020"},{"key":"B26","doi-asserted-by":"publisher","first-page":"735238","DOI":"10.1016\/j.aquaculture.2020.735238","article-title":"Occupational health, safety and work environments in norwegian fish farming-employee perspective","volume":"524","author":"Thorvaldsen","year":"2020","journal-title":"Aquaculture"},{"key":"B27","doi-asserted-by":"publisher","first-page":"1065","DOI":"10.1109\/JOE.2024.3492537","article-title":"Probabilistic particle filter anchoring (ppfa): a novel perspective in semantic world modeling for autonomous underwater vehicles with acoustic and optical exteroceptive sensors","volume":"50","author":"Topini","year":"2025","journal-title":"IEEE J. Ocean. Eng."},{"key":"B28","doi-asserted-by":"publisher","first-page":"e59187","DOI":"10.7554\/eLife.59187","article-title":"3deecelltracker, a deep learning-based pipeline for segmenting and tracking cells in 3d time lapse images","volume":"10","author":"Wen","year":"2021","journal-title":"Elife"},{"key":"B29","first-page":"10371","article-title":"Depth anything: unleashing the power of large-scale unlabeled data","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Yang","year":"2024"},{"key":"B30","first-page":"1","article-title":"ByteTrack: Multi-object Tracking by Associating Every Detection Box","volume-title":"Computer Vision \u2013 ECCV 2022. ECCV 2022. Lecture Notes in Computer Science, vol 13682","author":"Zhang","year":"2022"},{"key":"B31","doi-asserted-by":"publisher","first-page":"740429","DOI":"10.1016\/j.aquaculture.2023.740429","article-title":"Farmed atlantic salmon (Salmo Salar l.) avoid intrusive objects in cages: the influence of object shape, size and colour, and fish length","volume":"581","author":"Zhang","year":"2024","journal-title":"Aquaculture"},{"key":"B32","doi-asserted-by":"publisher","first-page":"104110","DOI":"10.1016\/j.apor.2024.104110","article-title":"Acoustic camera-based super-resolution reconstruction approach for underwater perception in low-visibility marine environments","volume":"150","author":"Zhou","year":"2024","journal-title":"Appl. Ocean Res."},{"key":"B33","doi-asserted-by":"publisher","first-page":"975","DOI":"10.1111\/raq.12218","article-title":"Intelligent feeding control methods in aquaculture with an emphasis on fish: a review","volume":"10","author":"Zhou","year":"2018","journal-title":"Rev. Aquac."}],"container-title":["Frontiers in Robotics and AI"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2025.1628213\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,17]],"date-time":"2025-10-17T07:41:12Z","timestamp":1760686872000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2025.1628213\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,17]]},"references-count":33,"alternative-id":["10.3389\/frobt.2025.1628213"],"URL":"https:\/\/doi.org\/10.3389\/frobt.2025.1628213","relation":{},"ISSN":["2296-9144"],"issn-type":[{"value":"2296-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,17]]},"article-number":"1628213"}}