A computer vision approach for the automatic detection of social interactions of dairy cows in automatic milking systems

The integration of digital technologies and Artificial Intelligence (DT&AI) in veterinary practice is one of the key topics to improve Herd Health Management (HHM). The HHM includes the prevention of diseases, the assessment of the welfare, and the sustainability production of farm animals. In dairy cattle farming, particular attention is paid to automatic cow detection and tracking, as such information is closely related to animal welfare and thus to possible health issues. Cows are highly social animals; therefore, a better comprehension of social context can help improve their management and welfare. In the field of Precision Livestock Farming, computer vision represents a suitable and non-invasive method for automatic cow detection and tracking. In this study, we developed and tested the reliability of a deep learning-based computer vision system for the automatic recognition of dairy cows in a barn equipped with Automatic Milking System. We aimed to build the social network of 240 dairy cows (primiparous and multiparous) to understand how social interactions can influence their welfare and productivity.