Swimming behavior as a potential metric to detect satiation levels of European seabass in marine cages

Abstract

Aquaculture is anticipated to contribute to two-thirds of the world’s fish consumption by 2030, emphasizing the need for innovative methods to optimize practices for economic viability, social responsibility, and environmental sustainability. Feeding practices play a pivotal role in aquaculture success and the feeding requirements are dynamic, influenced by factors like fish size, environmental conditions, and health status necessitating ongoing improvements in feeding practices. This study addresses a critical gap in feeding control systems in sea cages. It introduces a continuous, real-time monitoring system for analyzing the feeding behavior of European seabass, employing advanced AI models (YOLO and DEEPSORT) and computer vision techniques. The investigation focuses on key parameters, including speed and the newly defined feeding behavior index (FBI), to evaluate swimming responses under varying feeding scenarios exploring meal frequency, feeding time, and feeding quantity. The findings reveal a sensitivity of fish speed and the feeding behavior index (FBI) to different feeding scenarios, elucidating distinct behavioral patterns in response to varying frequencies, times, and quantities of feeding, such as increased activity in the morning relative to later times and the emergence of asymmetric activity patterns when fish are underfed or overfed. Notably, this study is one of the few in the field, presenting the development of a continuous, real-time monitoring system for feeding control in sea cages. Simultaneously, it explores reference curves and threshold values to enhance the overall efficacy of feeding control measures.