Abstract
Ensuring fish welfare is paramount in meeting the increasing demand for farmed salmon. Real-time monitoring is a key tool for gaining insights into fish behaviour, which is crucial for sustainable and ethical aquaculture practices.In this thesis, we explore current methods for monitoring fish behaviour in offshore salmon farming, and apply some of them, such as submerged cameras and single-ping echosounders to investigate salmon behaviour. We then develop a new and automated approach using machine vision techniques to detect fish and its behaviour patterns in 3D sonar data collected from a novel multi-beam echosounder infrastructure. Based on this method, we also identify and introduce some new parameters that can be extracted from 3D sonar data to further enhance our knowledge of fish behaviour. Finally, we assess how future far-offshore farming operations can communicate observations to expert systems on shore in more reliable and efficient ways.
Paper I: Reviews current and state-of-the-art fish farming observation methods.
Paper II: Investigates how environmental factors like waves and currents affect salmon behaviour and cage space.
Paper III: Explores the use of multibeam echosounders and machine learning for advanced monitoring.
Paper IV: 3D multibeam sonar is used to analyse the behaviour of offshore farmed salmon.
Paper V: Assesses the viability of cellular communication in remote aquaculture monitoring.
Together, these studies advance our understanding of monitoring technologies and salmon behaviour in aquaculture.
| Date of Award | 1 Mar 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Sponsors | Research Council Faroe Islands, Føroyagrunnurin, Fiskaaling - Aquaculture Research Station, P/F Firum & Waive AS |
| Supervisor | Qin Xin (Supervisor), Øystein Patursson (Supervisor) & John Potter (Supervisor) |
Keywords
- aquaculture