Automatic real-time monitoring of locomotion and posture behaviour of pregnant cows prior to calving using online image analysis

摘要: Monitoring the locomotion and posture behaviour of pregnant cows close to calving is essential in determining if there a need for human intervention assist parturition. In this study an automatic real-time monitoring technique described detail which allows identifying posturing prior calving. For purpose video surveillance images eight last 24h their were analysed. Data on seven different variables with time obtained each cow using monitor. These namely: x-y coordinates geometrical top view centre point cow; walking trajectory; distance walked; orientation main axis; body width/length ratio; hip length back area. then used classify specific behaviours such as standing or lying (including incidences motion during lying), eating drinking. On average 85% 87% drinking before could be correctly classified. However, developed needs further validated additional tests field.