Answers are grounded in this paper and cite their sources.
Investigates automated LED intervention for rapidly dispersing poultry piling, suggesting improvements in animal welfare.
Piling is an undesirable collective behavior in which poultry tightly aggregate for extended durations. Negative welfare consequences include stress, risk of heat stress, and suffocation, sometimes leading to death. In egg production facilities, piling is associated with a reduction in egg production. Ideally, piling should be quickly detected and safely dispersed with minimum disruption to poultry, and the attendant harms of piling prevented. With this aim in mind, we developed a smart intervention system based on overhead cameras and arrays of light-emitting diodes (LEDs). A neural network was trained to detect piling from real-time video of the aviary floor. Then during normal activities, piling was automatically detected and an intervention triggered. The intervention developed here consists of a sequence of illumination patterns generated from wall-mounted LED arrays. This study describes the intervention system and outcomes of a 70-day experiment in a poultry farm to assess the efficacy of the LED interventions for dispersing piling. We found that interventions caused rapid piling dispersion within two minutes for the initial 28 days, after which efficacy declined due to habituation. This is the first demonstration of a fully automated detection-intervention system tested for 70 days with hens.
No takes yet. Share an insight, caveat, or question.
Morris et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: