论文部分内容阅读
奶牛等大型动物的疾病和发情状况目前主要依赖饲养员目测判断,大规模集约化养殖仍采用人工观测方法,这不仅带来繁重的人力负担,也容易误判。为了能自动准确地识别奶牛是否发情或生病,该文提出在奶牛颈部安装无线传感器节点,通过各种传感器获取奶牛的体温、呼吸频率和运动加速度等参数,采用K-均值聚类算法对提取的各种参数进行行为特征多级分类识别,以此建立的动物行为监测系统能准确区分奶牛静止、慢走、爬跨等行为特征,从而可以长时间监测奶牛的健康状态。而且,这种监测系统易于推广到对其他动物的监测,对促进养殖业和畜牧业的发展也具有指导意义。
The diseases and estrous conditions of large animals, such as cows, depend mainly on the visual judgment of breeders. Large-scale intensive farming still uses manual observation methods, which not only brings heavy human burden but also is easy to misjudge. In order to automatically and accurately identify cow’s estrus or illness, this paper proposes to install wireless sensor nodes on the cow’s neck, obtain the cow’s body temperature, respiratory rate and acceleration through various sensors and other parameters. Using K-means clustering algorithm to extract , The behavioral characteristics of multi-level classification and identification of various parameters, animal behavior monitoring system can be established to accurately distinguish the behavior characteristics of cows at rest, walking, climbing, etc., which can monitor the health status of cows for a long time. Moreover, the ease with which this monitoring system can be extended to the monitoring of other animals is also instructive in promoting the development of aquaculture and livestock husbandry.