Behavioral Characteristics and Statistics-Based Imaging Techniques in the Assessment and Optimization of Tilapia Feeding in a Recirculating Aquaculture System

Optimization of feeding control for fish welfare is emphasized in aquaculture. To improve the feeding quality of tilapia in recirculating aquaculture systems (RAS), a novel method based mainly on shoal behavior for assessing the feeding of tilapia was proposed in this study. First, given the disturbances caused by negative factors, the foreground of the shoal was segmented using the average background modeling method in the YCbCr color model. Second, to avoid the complex tracking of individuals, optical flow was adopted to extract the behavioral characteristics (speed and turning angle) of the shoal. Third, with the help of statistical analysis, the obtained behavioral characteristics were standardized. Finally, regarding the digesta index of stomach and bowel (DISB), the feeding intensity of the shoal was assessed using combination entropy. In addition, two smart-feeding schemes combined with the feeding assessments of the shoal were developed. The results of this study demonstrated the feasibility of the proposed method for feeding assessments of tilapia with different DISB in three intervals (9.60 to 30.35, 104.61 to 126.84, and 282.73 to 293.43), as well as the feasibility of the smart-feeding schemes developed in this study.