A Study on Artificial Intelligence Techniques for Automatic Fish-Size Estimation

Biswas, R., Mutz, M., George, N., & Werth, D.(2023)
In: Intelligent Computing. SAI 2023, (Lecture Notes in Networks and Systems, vol 711), Hrsg: Springer, S. 1.116-1.126, DOI: https://doi.org/10.1007/978-3-031-37717-4_72

Link zur Publikation:


Abstract:

Understanding the distribution of body length or size of farmed fishes is a critical component for creating a sustainable and successful aqua-farming environment. This helps in gaining insight about population dynamics, average weight of fish stocks and reduces overfishing. Also, it helps in making informed business management decisions. Traditionally, the length of fishes are estimated manually. However, this approach is labor, time and cost intensive which makes it unscalable. Moreover, it leads to very poor population-level estimates. In this work, we conduct a survey of AI based techniques that can be applied successfully to resolve this problem. These approaches allow massive sampling of fishes in an automated way for improving fish size estimation at the population level. This survey is significant, important and helpful for aqua firming and particularly for food resource production. Our objective is to enhance and strengthen industrial growth for fish resource production through innovative farming. We discuss the different aspects, advantages and disadvantages of each approach. Finally, we compile our findings in a state-of-the-art survey on this topic.