Comparative Analysis of Biomass Estimation Methods in Aquaculture: A Review

Abstract

India is the third largest producer of fish in the world, behind China and Indonesia. Andhra Pradesh is the top state in India when it comes to fish output. Around 68 percent of the overall fish production is attributed to the aquaculture sector. Aquaculture contributes 1.07 percent to the nation's Gross Domestic Product (GDP). India is expected to use 1.6 million tonnes of fisheries by 2025. However, due to recent changes in local weather conditions, there has been a decrease in the productivity of aquatic ecosystems. The fish population in a concentrated aquaculture environment may provide useful data for the development of effective factory management systems. Advanced technologies play a vital role in augmenting the quality of products and raising the efficiency of production in aquaculture. Integrating automated fish identification technologies will allow precision farmers to attain enhanced efficiency and scientifically controlled output. Computer vision methods, a crucial domain of artificial intelligence, have emerged as a potent tool for automated fish detection. This is feasible because to the extensive utilisation and availability of contemporary information technology, including the internet of things, big data, and camera devices. Currently, it is often used to estimate biomass. Nonetheless, the creation of computer vision models for fish recognition encounters several obstacles, including varying illumination, diminished contrast, substantial noise, deformed fish forms, frequent blockage, and shifting backdrops. This work conducts a thorough investigation of biomass estimate for fish identification in several application situations.