Introduction
Image processing and computer vision technologies play a crucial role in fish detection, encompassing species identification, population estimation, and behavior analysis (Pennington et al., 2019 ; Yang et al., 2021) . While black soldier fly larvae meal (Hermetia illucens) (BSF) has gained traction as a dietary component in laboratory fish trials (Tran et al., 2022) , the behavior of fish fed different dietary BSF remains largely underexplored. This study investigated the feed preference of rainbow trout (Oncorhynchus mykiss) fed four diets containing 0%, 2.5%, 5%, and 10% dietary BSF meal using computer vision and open-source video analysis.
Materials and methods:
Four isoproteic (47%), isolipidic (17%), and isoenergic (22 MJ/kg) diets were formulated and extruded into 4 mm diameter pellets. Eighty-four rainbow trout (158.9 ± 3.0 g) were randomly assigned to twelve 50 L tanks (7 fish/tank) and fed one of the four diets (three replicates per diet) for five days. Twelve digital cameras were installed above each tank to record feeding activity. Feed intake, pellet count, and heatmaps depicting fish and pellet movement were generated from the video recordings.
Results
Heatmaps generated from video recordings revealed similar patterns of fish and pellet distribution across all dietary groups (Fig. 1). There were no significant differences in feed intake among the experimental diets. Fish in all groups consumed the delivered feed entirely within 25 seconds of feeding activation, and no significant variations were observed in pellet consumption time (Fig . 2).
There was no significant difference in pellet consumption time among the dietary groups (P = 0.19) (Fig. 3).
Interestingly, pellet consumption time was positively correlated with dietary BSF inclusion level (P = 0.042, Adjusted R-squared: 0.282 , F-statistic: 5.33) (Fig. 4 ). This suggests that fish took longer to consume diets containing BSF compared to the control diet.
Conclusion
This study demonstrates that despite comparable feed intake, fish fed BSF-containing diets exhibited a delayed feeding response compared to those fed the BSF-free diet. This suggests that BSF may be less palatable than fishmeal. Our findings propose an innovative, time-saving, and user-friendly approach to assess feed palatability in fish, surpassing conventional methods based solely on feed intake measurements. This approach holds potential implications for optimizing feed formulations and feeding practices.
Reference:
Pennington, Z.T., et al. ezTrack: An open-source video analysis pipeline for the investigation of animal behavior. Scientific Reports. 9, pp. 19979, (2019).
Tran, H.Q., et al. Systematic review and meta-analysis of production performance of aquaculture species fed dietary insect meals. Reviews in Aquaculture. 14, pp. 1637-1655, (2022).
Yang, L., et al. Computer Vision Models in Intelligent Aquaculture with Emphasis on Fish Detection and Behavior Analysis: A Review. Archives of Computational Methods in Engineering. 28, pp. 2785-2816, (2021).