基于超声波的母猪产前行为监测系统设计
Monitoring system design of sow's antenatal behavior based on ultrasonic

针对目前养殖业对母猪分娩时间的判断主要通过人工观察，不仅工作繁杂，而且易受饲养员主观经验影响等问题，设计一种基于超声波传感器和无线传感网络的母猪产前行为监测系统。该系统通过节点采集母猪产前头部、背部、尾部活动量的距离信息，将采集到的距离数据实时发送到网关节点，网关节点将距离信息转发到服务器。服务器端采用K-means聚类算法进行行为识别与分类。试验结果表明：系统能够快速采集母猪身体活动量的距离信息并对母猪行为进行分类，能够检测出母猪筑窝、站立、躺卧等行为，正确率为90.47%，系统工作稳定。提出了一种非接触式监测母猪产前行为的方法，为母猪分娩时间的预测提供基础。
The time of farrowing judgment usually relies on human observation,not only causes work complications,but also leads to problems by keepers own subjective experience.A system to monitor the behavior of sow is designed based on ultrasonic sensors and wireless sensor networks.Nodes of the system collect the distant information of head,back and tail of the sow,and transmit data obtained to gateway node in real time.The distant information is then transmitted to server from the gateway.Server uses K-means clustering algorithm to identify and classify saw's behavior.Testing results confirm that the system could quickly collect sow's physical activity,classify its behavior and detect its nesting,lying and standing behavior.The average correct rate is 90.47%,and the performances of the system are stable during the test.A non-contact method of monitoring the behavior of sows is proposed and provides a basis for predicting the time of farrowing.