Tracking and analysing social interactions in dairy cattle with real-time locating system and machine learning

There is a need for reliable and efficient methods monitoring the activity social behaviour in cows, order to optimise management modern dairy farms. This research presents an embedded system that could track individual cows using Ultra-wideband technology. At same time, interactions between individuals around feeding area were analysed with computer vision module. Detections of cows’ negative positive performed on foreground video stream Long-term Recurrent Convolution Networks model. The sensor fusion was implemented tested seven during 45 days experimental farm. performance evaluated at area. real-time locating based technology reached accuracy mean error 0.39 m standard deviation 0.62 m. detecting affiliative agonistic 93.2%. study demonstrates potential cows.