Advances in fish tracking technologies for aquaculture: Overcoming challenges and shaping future research

Aquaculture is essential for developing countries to have food security and for fishermen's socioeconomic conditions to improve. Fish tracking is essential to intelligent fish farming since it helps with health assessments, behavior monitoring, and water quality maintenance. However, high individual resemblance, rapid movement, and occlusions from foam in tanks provide obstacles for multi-object fish tracking. This paper explores the level of digital technology in aquaculture today, emphasizing systems based on vision, acoustics, and biosensors. It draws attention to the benefits, drawbacks, and uses of various technologies while highlighting important areas that require more study. Development is still hampered by a lack of extensive fish datasets and standardized evaluation techniques. We outline future research possibilities and move into advanced deep learning techniques like tracking-by -detection and merging deep features with correlation filtering to address these issues. We also give an overview of pre-deep learning fish tracking systems. This review provides a comprehensive overview of the evolution of fish tracking technologies and outlines potential avenues for advancing research and technology in aquaculture.