Feeding is a crucial aspect of fish farming in aquaculture that has a significant impact on fish health and growth. There are two categories of feeding practices in aquaculture, including commercial and non-commercial feeds. Commercial feeds are formulated using high-quality ingredients and are balanced in their nutrient content, while non-commercial feeds are made from locally available materials and are cheaper but often lack essential nutrients.
The feeding practices in aquaculture are dependent on factors such as fish species, age, size, and environmental conditions. Proper feeding practices are crucial to ensure healthy growth and development of fish, as overfeeding can lead to water pollution and reduced oxygen levels, while underfeeding can cause stunted growth and low survival rates. Therefore, it is essential to implement appropriate feeding practices to ensure the optimal growth and health of fish in aquaculture.
Role of Aquaculture in Human Life
Aquaculture plays a significant role in human life as fish make up more than half of all invertebrate species, and there are about 32000 different species of fish in the world, occupying almost every type of underwater habitat. Fish provide 15.7% of the protein consumed by the world’s population. With the development of nutritional treatments, aquaculture can produce healthy and affordable fish and shellfish to meet the growing demand.
The majority of fish consumed globally are predicted to be produced in aquaculture by 2030, making fish feeding and management essential in this industry. Conventional fish-feeding operations have high operating costs, low worker productivity, and high environmental pressure, making it important to explore new technologies such as computer vision-based systems for efficient and sustainable feeding.
Fishing is a significant contributor to the economy of Pakistan, providing a crucial source of income for those living along the coastline. In addition to sea fisheries, inland fisheries are also a very significant activity throughout the nation. Fish has a great nutritional value thanks to its 15–20% protein content, low cholesterol, and a variety of beneficial health additives.
In 2004, a total of 225 million tons of fish and fisheries products were exported, bringing in Rs. 7.6 billion. In comparison to 2003–2004, when the total marine and inland fish product was expected to be 566200M tons, the total marine and inland fish product for 2004–2005 was estimated to be 600M tons, of which 500M tons were marine product and the remaining 100M tons came from inland waterways. The study at hand is an effort by researchers to assess the profitability of the fishing business in Northern Regions and to estimate the overall cost of the structure for marketable trout fish husbandry.
What is a computer vision system in Aquaculture?
Computer vision system in aquaculture refers to an advanced technology that is used for fish feeding and monitoring fish behaviour. This system is highly automated and helps in ensuring better growth and quality of fish. The use of intelligent feeding systems is gaining importance in the aquaculture industry as it helps in increasing the productivity and quality of fish.
China, which is one of the largest mono-culture nations in the world, is also adopting these technologies to improve efficiency and productivity. The integration of advanced information technology such as smart vision and camera bias is helping in the development of intelligent products and choices in the aquaculture industry.
Artificial intelligence, especially computer vision technology, is playing a crucial role in observation and decision-making in aquaculture. Digital cameras with 2D and 3D vision systems are employed to gather data, which is then analyzed semantically by computers to make informed decisions.
The role of computer vision in fish feeding
Vanacloig has developed a reliable dynamic reactive 2D design utilizing computer vision to ensure that only tuna is extracted during aquaculture. This is achieved by using a deformable tuna model that is suitable for fish and is based on a stereo vision system and fish belly side profile. Accurate selected fish identification is crucial for aquaculture, and target detection can be used to track specific fish or populations.
Selected tracking detection is also used to analyze fish behaviour, identify goals, and detect abnormal behaviour. The preprocessing stage includes improving and de-noising underwater photos, which is a common finding in studies on fish target identification and underwater picture enhancement.
Observe Technologies offers a draw-and-play AI and data-processing solution to track quantifiable trends in feeding and provide growers with scientific advice on the significance of feeding. e-Fishery has developed a system that uses detectors to identify starving shrimp and fish, managing dispensers to provide appropriate amounts of food, effectively reducing nutrition expenditures by over twenty-one per cent.
Umitron Cell, a distant Japanese and Singaporean mono-culture technology company, provides a confluent of smart fish that is always controllable. This system optimizes feeding schedules and provides data-driven decision-making assistance, reducing waste, improving profitability and sustainability, and offering a better work-life balance for producers by eliminating the need to be out in the water in dangerous conditions.
Types of water quality detection robots
There are different types of water quality detection robots used in aquaculture to monitor and maintain optimal water conditions for fish growth. Underwater robots use location and guidance, path monitoring, and path planning critical technologies to identify real-time 3-D water quality. They are capable of real-time and stereoscopic water quality surroundings monitoring. Water surface unmanned ships can achieve three-dimensional water quality detection by raising water quality instruments.
Luna has developed an autonomous ship system that efficiently gathers and stores data on various water quality parameters such as ammonia nitrogen, pH, dissolved oxygen, turbidity, and conductivity. This system improves production efficiency and lowers labour risk while providing real-time monitoring of the aquaculture environment. A robot for tracking water quality in aquaculture has also been developed to increase productivity.
Karimanzira has developed autonomous underwater vehicles for monitoring, inspecting, and testing fish behaviour and water purity. These vehicles operate on the principles of route planning, obstruction avoidance, underwater positioning, and motion control. Frequent, economical, and lasting water quality testing is required for this type of monitoring. Devices are used to instantly determine water conditions in vast sea regions at various depths.
Image processing tools can be used to determine and identify the size of various fish species in aquaculture. A method has been developed using numerous fish species, including Hippoglossoides platessoides, Solea vulgaris, Microstomus kitt, Pleuronectes platessa, Sebastes marinus, Sebastes mentella, and Platichthys flesus.
Factors Influencing Fish Feeding
Fish feeding and appetite are influenced by various internal and external factors. The investigation of these factors can aid in the development of numerous intelligent feeding management techniques and systems. The internal factors that impact nutrition consumption and appetite in fish are their physiological and nutritional requirements. Ronnestad has found that the nervous and endocrine systems play a role in controlling fish feeding habits.
The physiological modulation of hunger is closely linked to peripheral and central impulses from the brain. Research has suggested that fish appetite, which is reflected in fish growth, can be directly influenced by nutrients. The external factors that impact fish feeding include the environment and managerial practices. The most significant environmental factors are temperature, photoperiod, and dissolved oxygen.
In some species, feeding activity is strongly correlated with temperature or photoperiod, and feed intake rates increase at higher temperatures and dissolved oxygen concentrations. Salinity may also impact the development and feed assimilation rate of some fish species. Understanding these factors is critical to developing effective feeding management practices in aquaculture.
Feeding control methods based on computer vision
Using a single camera to obtain data has clear benefits, and studies have shown that the area of a fish’s picture is closely linked to its mass. For example, one can estimate the weights of grey mullet, carp, and tilapia based on their pixel area. Researchers have demonstrated that it is possible to autonomously provide feed by using machine vision to count and determine the number of fish.
A single camera-based system has been developed to measure behavioural differences at each feeding stage with minimum frame loss, providing data for intelligent feeding. When the leftover feed seen by cameras hits a certain level, feeding is halted. Cameras are often used to provide input to determine feeding conditions.
Installing a leftover bait quantity calculator and a residual feed collection device, which can halt feeding when the residual bait amount hits a certain level when used with a camera, will enable intelligent feeding. Similar business devices are currently being used for cage feeding.
CONCLUSION
In recent years, there has been significant growth in the use of effective, automated, and precise smart aquaculture in the aquaculture sector. Researchers and aquaculture farmers have explored the application of artificial intelligence with programs or machine vision in efficient aquaculture.
The most widely used applications of this technology are in fish disease assessment, fish mounting, categorization and identification, nutrition management, and water state tracking. If this technology is used in Pakistan’s aquaculture, it can lead to a good harvest and be beneficial to the country’s economy. This technique is not too expensive, and it can save on labour costs while providing a good-quality harvest of fish culture.
Umm E Maria, Malaika Eman
University of Agriculture Faisalabad
