Deep Learning Technologies for Monitoring Poultry Wellbeing in the USA
编号:3
访问权限:仅限参会人
更新:2025-04-26 16:47:33 浏览:19次
特邀报告
摘要
The United States is currently the world’s largest producer of broiler chickens (9-10 billion chickens) and the second-largest producer of eggs (110 billion table eggs annually). Poultry production contributes $60 billion in farm gate value each year. However, there are ongoing concerns regarding animal welfare, particularly with fast-growing broiler chickens and the increasing use of cage-free housing for laying hens. Approximately 30% of fast-growing broilers experience leg problems and 50% or higher birds have footpad lesion. For egg production, cage-free hens face doubled mortality and injury rates compared to conventional caged systems. Therefore, precision farming technologies are needed to address those issues for sustainable production. To address these challenges, artificial intelligence (AI) technologies such as machine vision and robotics were innovated for monitoring poultry health and welfare in this study. The performance of those deep technologies has reached 90% or higher in monitoring footpad dermatitis, bumble foot, and mortality, etc. In addition, large foundation models such as SAM (Segment Anything Model) and TAM (Track Anything Model) were tested efficient in individual poultry tracking and locomotion speed monitoring.
关键词
Poultry production; animal welfare; machine vision; robots.
稿件作者
Lilong Chai
University of Georgia
发表评论