The cultivation of fish and shellfish stands to benefit significantly from the integration of artificial intelligence techniques and advanced technologies, both in terms of enhancing practices and boosting sustainability
The FAO has documented that the application of AI to aquaculture, a practice developed over four millennia, has the potential to substantially improve various aspects of this field. Presently, aquaculture is responsible for producing nearly half of the fish products consumed globally, with demand continually on the rise. It ranks as one of the fastest-expanding sectors in global food production and plays a significant role in the worldwide food supply chain and economic development.
The International Trade Administration projects that the global aquaculture market, currently valued at 204 billion dollars, is poised to reach 262 billion dollars by the close of 2026.
Beyond its economic significance, aquaculture’s development must prioritize sustainability. Its inclusion among the 17 goals of the 2030 Agenda underscores its importance; furthermore, in terms of sustainability, the management of fisheries and aquaculture is a key facet of the blue economy.
AI, along with related technologies like robotics and drones, is invaluable in enhancing and sustaining aquaculture practices. These technologies offer diverse applications and benefits: AI systems can monitor water quality, temperature, and overall conditions of the species being cultivated.
Additionally, robots and drones can perform comprehensive checks and maintain aquaculture infrastructure.
However, this field is just at its nascent stage; a myriad of applications is poised to enhance the wellbeing of farmed animals and species, ensuring optimal environmental conditions for their habitation.
Takeaways
Understanding Aquaculture
Aquaculture involves the cultivation, management, and harvesting of fish, crustaceans, and aquatic plants. An age-old practice, it was undertaken by ancient civilizations such as the Egyptians and the Chinese. Today, it continues to be a vital system for producing food and other commercial products, restoring habitats, and replenishing wild stocks, thereby aiding in the preservation and growth of threatened or endangered fish and aquatic species.
With an escalating demand for fish products, technology has enabled the cultivation of food in both coastal marine waters and the open sea, as well as in freshwater environments, turning it into a source of both environmental and economic wealth. According to the FAO, global aquaculture fish production in 2018 reached a record high of 82.1 million tonnes, including 32.4 million tonnes of aquatic algae, and 26 thousand tonnes of ornamental shells and pearls cumulating in a historic total of 114.5 million tonnes. In that year, aquatic animal farming was predominantly focused on fish, accounting for over 54 million tonnes, valued at nearly 140 billion dollars.
In Europe, aquaculture is crucial in coastal and riverine areas, with the sector in 2020 achieving 1.2 million tonnes in sales volume and a turnover of 3.9 billion euros. It provides direct employment to approximately 57,000 people across around 14,000 predominantly small, family-run businesses.
67% of the EU’s aquaculture output is concentrated in Italy, France, Greece, and Spain.
The European Commission highlights that
“sustainable development of aquaculture is a principal goal of the common fisheries policy. Furthermore, the European Green Deal recognizes aquaculture production as a source of ‘low-carbon’ proteins for food and feed”.
To this end, the application of AI in aquaculture can make a significant contribution, particularly in assessing, monitoring, and mitigating the environmental impact of aquaculture activities.
Artificial Intelligence for Aquaculture: Case Studies
The integration of algorithms and AI methods in aquaculture capitalizes on the substantial and multi-tiered advancement of technology. This includes the Internet of Things, cloud computing, and 5G networks, which have revolutionized data collection, circulation, and processing at previously unattainable speeds.
AI is increasingly being evaluated and implemented in aquaculture to enhance feed efficiency, biomass estimation, growth monitoring, early disease diagnosis, environmental monitoring and control, and to reduce labor costs. With the advent of modern sensing and processing technologies, many routine aquaculture tasks can now be automated, improving animal welfare conditions.
For instance, consider deep learning: a research team led by Professor James C. Chen from the National Tsing-Hua University in Taiwan utilized a pre-trained deep learning model for the visual analysis of a grouper farm. The advanced identification of fish appearances or abnormal conditions underwater, achieved through computational intelligence and the adoption of isolation measures, facilitated anomaly detection and reduced the risk of pathological contagion among fish.
The experiment demonstrated that the pre-trained model could classify three different types of abnormal grouper appearances with an average accuracy of nearly 99%.
Similarly, a research team coordinated by Professor Narayana Darapaneni, an Artificial Intelligence and Machine Learning lecturer at the University of Bangalore in India, developed a system for the early detection of disease outbreaks, aimed at assisting small fish farmers. This system, based on underwater cameras or sensors, transmits images via the cloud for processing. Once classified and analysed by a trained AI model, modern connectivity options can reduce data processing times to a few minutes, enabling multiple farm evaluations daily.
Robotics in Aquaculture of Fish and Shellfish
In concert with artificial intelligence in aquaculture, robotics plays a pivotal role in fostering sustainable development. Notably, Canada’s Ocean Supercluster has unveiled the ARCAP (Autonomous Robotic Capabilities for Aquaculture Project), a 2-million-dollar venture focused on establishing a comprehensive autonomous system within the aquaculture sector. This initiative utilizes autonomous underwater vehicles and surface vessels, furnishing invaluable data on fish health, population size, and dimensions, alongside insights into infrastructure and water quality.
In addition, the “Visual Assessment of Aquaculture Pens Project,” funded with 150,000 dollars, is set to commence. This endeavour will see the development of visual assessment tools for the autonomous underwater vehicle Aqua2, facilitating independent observation, surveillance, and appraisal of aquaculture net pens and their fish inhabitants.
Europe, too, is making strides in aquaculture robotics, with the Norwegian company Remora Robotics at the forefront. They have engineered a robot, powered solely by electricity, designed for cleaning and inspecting fish enclosures whilst concurrently gathering pertinent data. This innovation is especially adept at preventing biofouling – the undesirable accumulation of microorganisms, plants, algae on maritime structures. Continuous monitoring helps avert potential net irregularities. Equipped with environmental sensors, the robot can conduct analyses and relay comprehensive information on net conditions and their progression over time, thus enabling proactive maintenance.
Clean nets ensure better water circulation, elevated oxygen levels, and suitable temperatures – all contributing to enhanced water quality and living conditions for fish.
The Future of Precision Aquaculture: Drones, Lidar, and Swarm Robotics
The integration of AI in aquaculture paves the way for a future where digital technology is key to enhancing production and improving the living conditions of farmed species, as well as their environmental habitat. AI’s role extends beyond monitoring and data analysis of water quality, fish health, and environmental conditions; it also encompasses the development of swarm robotics. This emerging technology involves autonomous robots collaborating towards a shared objective. Within aquaculture, such robots could be deployed for water quality management, disease detection, and production optimization. They may also revolutionize the harvesting process, diminishing labor costs and amplifying efficiency.
Israeli startup GoSmart exemplifies this technological evolution, employing AI and machine learning to revolutionize fish farming. GoSmart has devised fully autonomous, energy-efficient systems that are compact, attachable to cages, ponds, or aquaculture tanks. Operated via an edge AI platform, these systems scrutinize the average weight and distribution of fish populations in their habitat, as well as temperature and oxygen levels. This data is then delivered through GoSmart’s software-as-a-service, aiding aquaculturists in making precise and efficient decisions regarding feeding and harvesting schedules in real time. Presently, GoSmart is enhancing its systems to analyze fish behavior and disease markers, augmenting its existing capabilities of assessing fish weight, population distribution, temperature, and oxygen levels.
Drones too have emerged as invaluable assets in aquaculture. Equipped with cameras and sensors, they can oversee aquaculture installations from above, gauging water quality parameters like temperature, pH, dissolved oxygen, and turbidity. Beyond surveillance, these drones can be adapted to methodically dispense feed, refining feeding practices.
Armed with cameras and computer vision technology, they play a crucial role in managing the environment, curbing the spread of invasive species, monitoring climatic conditions, identifying potential pollution sources, and evaluating the ecological impact of aquaculture activities.
Early detection of disease outbreaks is crucial in aquaculture. Drones with thermal imaging capabilities can pinpoint water temperature variations, serving as indicators of potential health issues. Additionally, they are instrumental in deterring birds and other pests, safeguarding aquaculture from potential hazards. LIDAR (Light Detection and Ranging) technology, often used in aerial surveys, equips drones with lasers to measure distances and create intricate 3D maps of the terrain, or in this case, water bodies. This technology offers non-intrusive, cost-effective means to gather accurate, real-time data on fish populations.
