Back to search

HAVBRUK2-Stort program for havbruksforskning

Bio-metric profiling of aquaculture enclosures utilizing hydro-acoustic sensors

Alternative title: Biometrisk kartlegging av oppdrettsmerde ved bruk av hydroakustikk

Awarded: NOK 3.7 mill.

Project Number:

327732

Project Period:

2021 - 2023

Introduction This project was initiated with the objective of advancing aquaculture management through the development of hydroacoustic technology. The focus was on enabling fish farmers to monitor and manage fish biomass in their pens more effectively. This technology aimed at providing data on various parameters such as individual fish size, average size, overall size distribution, and the dynamics of biomass over time and spatial dimensions. The intention was to utilize advanced hydroacoustic methods to offer a more precise and real-time monitoring solution. Project Implementation A collaborative effort between Furuno and NTNU formed the backbone of the project. Resources were allocated from both entities, involving two full-time employees and several part-time contributors. The team composition included a lead researcher from Furuno, a post-doc from NTNU and additional supporting resources from Furuno. In addition, Furuno supported through the project several related studies performed by NTNU and NOFIMA. While the project exceeded its initial budget, the additional time and costs were absorbed by the participating entities, highlighting a significant investment in research and development. Objectives The project was guided by a set of primary and secondary objectives: Primary Objectives: • To develop a hydroacoustic solution capable of accurately determining fish weight distribution in pens, aiming for a 2% accuracy level on average weight and distribution, incorporating short measurement cycles or real-time scanning. • To create algorithms using machine learning for analyzing biomass dynamics, focusing on enhancing weight distribution assessments. • To assess the potential of machine learning and hydroacoustic data in determining critical aspects of fish health, such as appetite and sexual maturity. Secondary Objectives: • To collect 2D/3D data to enable research capabilities beyond current standards. • To improve sampling methodologies for better reference points in studies. • To establish academic research methods to enhance result validation. • To integrate biological models, particularly based on Atlantic Salmon, into hydroacoustic algorithms. Methods The project addressed three main challenges: Hydroacoustic Accuracy Evaluation: Collaborative experiments with NTNU focused on assessing the accuracy of hydroacoustic in estimating individual fish weight. These experiments included controlled tests in smaller settings and broader analyses in production pens. Ensuring accurate representation in sampling methods was a critical aspect of this challenge. Representative Sampling: Addressing the hypothesis that conventional camera systems might miss comprehensive coverage due to limited detection ranges, hydroacoustic equipment was employed to scan larger pen cross-sections. This approach aimed to provide a more accurate population representation. Biological Variation: Acknowledging the influence of biological variation on inferred weight, the project emphasized identifying and prioritizing biological factors that significantly affect weight calculations. This part of the study included a detailed literature review and collaboration with experts. Key Studies Key studies conducted during the project included: • Literature Review: Undertaken by NTNU, focusing on existing aquaculture research. • Single Fish Study: Conducted by NTNU to investigate individual fish characteristics. • WASSP Study: A joint effort exploring the use of Wide Angle Sonar Seafloor Profiler technology. • Morphology Study: NTNU-led study examining physical attributes of fish. • Sampling Methodology Study: Focused on refining sampling techniques for enhanced accuracy. • Shadowing Study: Part of the “European Study Group with Industry,” addressing sampling challenges in densely populated pens. • Weight Analysis: Comparing hydroacoustic measurements with traditional methods. • Behaviour Analysis: Conducted by NOFIMA to understand behavioural patterns. Results The project yielded several key outcomes: • Weight Estimation Accuracy: We achieved the targeted 2% accuracy in average weight estimation against traditional grader report data, incorporating real-time scanning and behaviour analysis within a day. • Biomass Scanning: Algorithms were developed to scan a larger vertical section of biomass, with ongoing efforts to extend this range, especially concerning the shadowing effect. • Behavioural Pattern Recognition: Both single beam and multibeam transducers were effective in identifying multiple behavioural traits. These findings are expected to be integrated into future commercial products.

Benefit for business • We have developed a commercial product able to detect average weight and weight distribution. We have increased staff and will from 2024 also employ sales personnel to launch product to market. • We have also established capability to further develop product to address behavior and health. • We have as an organisation build knowledge and understanding on a deeper level on our own products, but more importantly domain knowledge in aquaculture. This would enable the business to expand to a new area of business. • We have built capabilities to expand our business outside of Norway. Benefit for Customer / Society • With our product we provide deeper insight in the biomass dynamics and increase accuracy in bio-metric measurement, enabling the farmer to improve feeding operations, production planning, and address fish welfare. • It will reduce the workload for farmers with its capabilities and its low maintenance profile. • Through better understanding of biomass, logistics operations can be optimized and total environmental impact can be reduced. • Through a better understanding of biomass dynamics, insight in fish welfare can be improved. This will enable the fish farmer to better address issues in industry with fish health and mortality.

Current technology for analysis of biomass is largely base on spotlight measurement of individuals, without taking the natural variation and biomass dynamics into the equation. This means that the fish needs to swim past a sensor, and the profiles is generated from a time series of this measurement. The goal with this project is to develop a technology to provide analysis based on real time measurements at population level. With this technology one should be able to better understand the real time dynamic of the biomass, and thus gain insight in fish weight, behaviour in and surrounding the pen, appetite, welfare, health and maturation. To exploit existing knowledge both from fish farming companies, NTNU and Furuno Norway we aim at building causal diagrams to better establish statistical models to explain variation in data. Through causal inference with multiple covariates we aim to build prediction-models based on new insight from datasets not available with current technology, combined with extensive domain knowledge both in the domain of fish biometrics (Furuno) and the biology of atlantic salmon (NTNU) To develop this technology we aim at utilizing statistical multi model inference, causal diagrams and exploratory data analysis techniques combined with machine learning. To be able to develop this technology a better understanding of biological parameters is crucial for the accuracy of the product. Research is therefore needed to develop biological models taking into consideration the fish behavior and its biological variation. This research is intended to be done in close cooperation with NTNU.

Funding scheme:

HAVBRUK2-Stort program for havbruksforskning