FREC chart: A computer vision framework for quantifying feeding and foraging rhythms in beef cattle

Feeding management is critical in beef cattle farming, and understanding feeding events and cattle foraging rhythms is essential for improving feeding management and animal health. To support scientific and timely feeding management decisions, this study established a computer vision framework that constructed a digital index and a novel visualisation chart for expressing cattle feeding events, foraging behaviour, and ambient conditions. Specifically, beef cattle and their postures were detected from feeding alley video streams using YOLO-V8 model, and foraging behaviour was identified through geometric computation. The Average Foraging Rate ( AFR ) was calculated across sampling intervals, and the 24-hour period was divided into Typical Foraging Periods (TFPs), Atypical Foraging Periods (AFPs), and Non-Foraging Periods (NFPs). Hourly data of feeding management, AFR , and Temperature-Humidity Index (THI) were visualised through the Feeding-Rhythm-Environment Clock (FREC) chart. Seven days of continuous video data from two cattle farms were selected to validate the framework. The results showed that the cattle detection model achieved F1-scores of 84.1% and 80.2% at Farm A and Farm B, respectively. The posture classification model achieved F1-scores above 95.0% for standing and lying at both farms. The AFR during TFPs was significantly higher than that during AFPs ( p < 0.001), indicating its sensitivity to detect the group-foraging behaviour induced by feeding events and its validity for rhythm analysis. The proposed framework successfully identified management-related foraging rhythm disruptions and energy imbalances in Farm A and stable foraging rhythms in Farm B, which advanced behavioural health assessment and precision feeding management in beef cattle farming.