Characterizing locomotor behavior variability in commercial broiler flocks using large-scale video tracking

The surveillance of broiler chicken welfare requires objective and scalable approaches capable of capturing behavioral dynamics under commercial production conditions. This study characterized locomotor behavior in broiler flocks reared under commercial conditions using large-scale video observations. A multi-camera system was deployed across four commercial farms, encompassing 12 flocks and more than 14,500 h of recordings. Broilers were automatically detected and tracked to extract spatiotemporal movement descriptors. Three complementary indicators were analyzed: the trajectory segment-based proportion of inactivity, the mean speed of active trajectory segments, and collective movement events reflecting transient synchronization of motion. Statistical models were used to estimate expected temporal trajectories and prediction intervals for each indicator across age, daily activity period, sex, and farm. The results revealed structured patterns of locomotor organization across production cycles and farms, allowing the expected range of behavioral variability under routine commercial conditions to be quantified. When applied to observations recorded under non-routine conditions, deviations from these reference ranges occurred more frequently, indicating departures from typical flock locomotor dynamics. These findings demonstrate that large-scale video observations can provide quantitative reference patterns for broiler locomotor behavior under commercial conditions. Establishing such baselines offers a practical foundation for behavioral monitoring approaches aimed at identifying atypical flock activity patterns in commercial poultry production systems.