Sensor-based lameness detection in group housed sows

欧盟第2001/88/EG号指令规定了怀孕母猪的集体饲养；原因是与畜栏饲养制度相比，这是一种对动物更友好的程序。然而，在母猪分组后，群体住房中更容易出现跛行，这在很大程度上取决于地板设计的频率和严重程度、管理和动物行为(例如，竞技性互动)。跛行限制了母猪的福利，直接导致经济损失(例如兽医成本、屠宰收入)，并间接影响母猪的寿命和繁殖性能。此外，在生猪数量不断增加的群体中，监测单个猪的难度也很大。因此，该项目旨在开发群养母猪跛行的计算机辅助早期检测。将首次使用系列数据(定位、加速度计)建立监测系统，以支持人员的监测活动，并实现早期诊断和有针对性的治疗。定位数据，例如每头母猪每天行进的距离，将提供动物特有的但相当粗略的运动模式。由于频谱的时间局部化，利用小波对加速值(信号)进行分解将允许更准确地分辨信号，因此期望更好地对跛行进行早期警告。

EU Directive 2001/88/EG stipulates the group housing of pregnant sows; the reason being that it is a more animal friendly procedure compared to stall housing systems. However, lameness occurs more often in group housing immediately after the grouping of the sows, which depends significantly on the frequency and severity of the floor design, the management and the animals behaviour (e.g. agonistic interactions). Lameness restricts the welfare of the sows resulting directly in economic losses (e.g. veterinary costs, slaughter revenues) and indirectly influencing the longevity and fertility performance of the sows. Added to this is the difficulty of monitoring individual pigs in a group with increasing pig numbers. The project therefore aims to develop the computer aided early detection of lameness in group-housed sows. For the first time, serial data (localisation, accelerometer) will be used to set up a monitoring system which supports the monitoring activities of the personnel as well as enable early diagnosis and targeted treatment. The localisation data e.g. the distance travelled per sow per day will provide an animal specific yet rather rough movement pattern. A decomposition of the acceleration values (signals) with the aid of wavelets will allow a more exact resolution of the signals due to the temporal localisation of the frequency spectrum and thus better early warning of lameness is expected.