Evaluation of Kinect depth sensor to monitor laying hens activities in a pilot henhouse

摘要

As increasing scale of farms and the corresponding higher number of laying hens, it is increasingly difficultfor farmers to monitor their animals in a satisfactory way. Early detecting abnormal animal activities is helpful for farmers' fast response to the negative impact on animal welfare and daily management. In this study, we introduce an automatic, non-invasive methods for detecting poultry abnormal activities and occupation zones in a pilot henhouse by using Kinect depth sensor. A typical regionincluding eight Hy-Line brown laying hens was continuously monitored by a top-view Kinect during 49 experimental days. A mean prediction model (MPM) was proposed to monitor animal activities and occupation zone based on training and testing set data. Compared with recorded farmer logbook, this method can report abnormal activities with an average accuracy of 84.2%. Additionally, it was found that the flock showed regular changes of activity and occupation index in day and night. Each week had a similarchanging pattern. Finally, this method can provide farmers a useful tool to monitor animal activity in order to improve their animal welfare and realize high-efficient management in a 24-h day.