AI-Based System for Monitoring Laying Hen Behavior in Small-Scale Poultry Farms

Strategic Insights -

A recent study by researchers from Prairie View A&M University has developed an AI-based system for monitoring laying hen behavior using computer vision for small-scale poultry farms. The system, designed for small barn environments housing at most 10-15 chickens, uses an object detection model trained with a dataset of laying hen, feeder, and waterer objects. The system processes each detection per frame using bounding boxes and movement-based approximation identification, simplifying the tracking process while providing valuable behavior insights. According to the research, the system offers an efficient, low-cost solution for monitoring chicken feeding and drinking behaviors, supporting improved management and early health detection.

Key Takeaways:

  • The AI-based system uses computer vision to monitor laying hen behavior in small-scale poultry farms, addressing the lack of advanced monitoring tools in these settings.
  • The system was trained with a dataset of over 700 frames, annotated manually with different lighting, hen positions, and interaction angles with dispensers.
  • The developed system consists of an object detection model created on top of the YOLOv8 model, which achieved an average mean average precision (mAP@0.5) metric value of 91.5% and a detection accuracy of over 92%.
  • The system uses bounding boxes and movement-based approximation identification to simplify the tracking process without losing valuable behavior insights.
  • The AI-based system is designed for small barn environments housing at most 10-15 chickens, making it a feasible solution for small-scale poultry farms.
  • The system provides an efficient and low-cost solution for monitoring chicken feeding and drinking behaviors, supporting improved management and early health detection.
  • The research highlights the urgent need for intelligent, low-cost systems that can continuously and accurately monitor bird behavior in resource-limited farm settings.

Statistics:

  • Over 700 frames were annotated manually for high-quality labeled data, used for training the detection model.
  • The object detection model achieved an average mean average precision (mAP@0.5) metric value of 91.5%.
  • The system detected chicken feeding and drinking behaviors with a detection accuracy of over 92%.
  • Global poultry production expands annually, raising over 70 billion hens each year.

Sources:

  • Agriculture, 2025,15(18):1963. (Agriculture - http://www.mdpi.com/journal/agriculture)
  • Our news journalists report that additional information may be obtained by contacting Jill Italiya, Department of Computer Science, Prairie View A&M University, Prairie View, TX 77446, United States.