A Novel 3-DOF Robotic Gantry with Underwater Stereo Vision and AI for Automated, Continuous Observation of Tilapia Behavior in Intensive Aquaculture Systems
Journal of Integrated Engineering Sciences · 05.04.2025 · 1 (2) · 110-142
DOI: 10.63686/jies.v1i2033 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Integrated Engineering Sciences, online erschienen am 05.04.2025. Verfasst von Ali Humaid Mohammed Al Alawi und 1 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Continuous monitoring of fish behavior and health is essential for improving productivity, animal welfare, and management efficiency in precision aquaculture. Conventional fixed-camera monitoring systems are often limited by restricted observation coverage, blind regions, and the requirement for multiple cameras to monitor large aquaculture facilities. This paper presents an automated underwater stereo vision monitoring system that integrates a three-degree-of-freedom (3-DOF) Cartesian robotic gantry, a waterproof stereo camera, and an artificial intelligence (AI)-based image analysis framework for continuous fish behavior and health assessment. The robotic gantry automatically positions the stereo camera at predefined inspection locations, enabling systematic acquisition of synchronized underwater stereo images from multiple viewing angles and depths. The software framework combines underwater image enhancement, stereo vision, three-dimensional (3D) reconstruction, deep-learning-based fish detection, multi-object tracking, behavioral feature extraction, health assessment, and real-time visualization through an integrated monitoring dashboard. A laboratory-scale prototype was developed to validate the proposed system. Experimental evaluation demonstrated sub-millimeter robotic positioning repeatability (±0.48 mm), a stereo calibration error of 0.34 pixels, and a mean depth estimation error of 1.9%. The AI framework achieved a fish detection precision of 96.7%, recall of 95.8%, and an overall health assessment accuracy of 95.5%, while supporting near real-time monitoring at 22 frames/s. The reconstructed swimming trajectories successfully quantified behavioral indicators, including swimming speed, activity level, turning frequency, swimming depth, and spatial distribution, enabling discrimination between normal and abnormal behavioral patterns. These results demonstrate the feasibility of integrating robotic automation, underwater stereo vision, and artificial intelligence into a unified platform for continuous fish welfare monitoring and intelligent decision support in precision aquaculture.
Zitieren
APA
Alawi, A. H. M. A., & Asopa, S. (2025). A Novel 3-DOF Robotic Gantry with Underwater Stereo Vision and AI for Automated, Continuous Observation of Tilapia Behavior in Intensive Aquaculture Systems. Journal of Integrated Engineering Sciences, 1(2), 110-142. https://doi.org/10.63686/jies.v1i2033
Vancouver
Alawi AHMA, Asopa S. A Novel 3-DOF Robotic Gantry with Underwater Stereo Vision and AI for Automated, Continuous Observation of Tilapia Behavior in Intensive Aquaculture Systems. Journal of Integrated Engineering Sciences. 2025;1(2):110-142. doi:10.63686/jies.v1i2033
BibTeX
@article{Alawi2025,
author = {Alawi, Ali Humaid Mohammed Al and Asopa, Sneha},
title = {A Novel 3-DOF Robotic Gantry with Underwater Stereo Vision and AI for Automated, Continuous Observation of Tilapia Behavior in Intensive Aquaculture Systems},
journal = {Journal of Integrated Engineering Sciences},
year = {2025},
volume = {1},
number = {2},
pages = {110--142},
doi = {10.63686/jies.v1i2033}
}
EUVIMED