A scene-adapted lightweight YOLO-LD model for dead fish detection in recirculating aquaculture systems

Key Points

• The YOLO-LD model achieves high accuracy in dead fish detection using advanced computer vision techniques.
• Evaluation shows that the model has a detection precision of over 90% in various aquaculture settings.
• Observational analysis across different light conditions reveals consistent performance of the detection model.
• This method may improve monitoring efficiency in aquaculture, suggesting reduced operational workload for farmers.