Scholarlink

<p>In poultry industry, the manual observation is the method used for traditional chicken health monitoring, which is time-consuming and prone to mistakes, especially at large- scale farms. This research suggests an automated real-time method for classifying chickens into three groups: Healthy, Unhealthy, and Not a Chicken. Because of its real-time detection capabilities, You Only Look Once (YOLOv5) model is used for objects detection. This automated system is integrated with Chicken swarm optimization (CSO) algorithm for monitoring of chicken health parameters thus enhancing the accuracy of the proposed approach. Both picture and video feeds are used as input for pre-processing followed by saving of real-time annotated visuals with bounding boxes and health status labels using CSO. CSO offers a scalable and affordable solution for automatic monitoring, thus reducing the spread of illness, and boosting productivity. The results show that the proposed approach achieves the highest accuracy (96.58%) thereby surpassing existing studies and ensuring effective poultry management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel approach for automatic monitoring and classification of chickens using YOLOv5 and CSO

  • 期刊论文

  • 发表日期：

    2025年8月20日

查看全文
  • Spoorthi P. Shetty,
  • Mangala Shetty,
  • Bharath S. Bhat,
  • Kavana Navada,
  • Kanthashree Jain

摘要

In poultry industry, the manual observation is the method used for traditional chicken health monitoring, which is time-consuming and prone to mistakes, especially at large- scale farms. This research suggests an automated real-time method for classifying chickens into three groups: Healthy, Unhealthy, and Not a Chicken. Because of its real-time detection capabilities, You Only Look Once (YOLOv5) model is used for objects detection. This automated system is integrated with Chicken swarm optimization (CSO) algorithm for monitoring of chicken health parameters thus enhancing the accuracy of the proposed approach. Both picture and video feeds are used as input for pre-processing followed by saving of real-time annotated visuals with bounding boxes and health status labels using CSO. CSO offers a scalable and affordable solution for automatic monitoring, thus reducing the spread of illness, and boosting productivity. The results show that the proposed approach achieves the highest accuracy (96.58%) thereby surpassing existing studies and ensuring effective poultry management.