| Abstract |
In modern livestock farming, dairy cows′ health and physical condition are crucial for milk production and its impact on the dairy industry. However, in subtropical regions, the high temperature and humidity environment pose different challenges for dairy cattle feeding and management. Due to environmental, space, cost, and technology limitations, comprehensive control becomes difficult, leading to various feeding management issues. These problems are closely related to dairy cows′ health status and body condition scoring. Traditional dairy cow body condition scoring mainly relies on trained inspectors, but this method leads to variability in scoring due to the subjective judgment and experience of the inspectors. Due to subjectivity, the repeatability of scoring is poor. Even the same inspector may vary in their assessments of the same cow at different times. Manual scoring requires a great deal of time and labor, especially in large farms where a large number of dairy cows need to be evaluated, leading to reduced efficiency. Therefore, this project plans to develop automated image analysis technology to explore the differences in subjectivity and objectivity in body condition measurement to solve the problems of manual measurement. In addition, by formulating a feeding management plan and using machine learning technology to predict the trend of changes in cattle body condition, effective prevention of cattle health problems and production efficiency can be achieved. |
