| 基于加速度传感器的种公羊运动行为识别 |
| |
| 引用本文： | 张曦宇,武佩,宣传忠,杨建宁,刘艳秋,郝敏.基于加速度传感器的种公羊运动行为识别[J].中国农业大学学报,2018,23(11):104-114. |
| |
| 作者姓名： | 张曦宇 武佩 宣传忠 杨建宁 刘艳秋 郝敏 |
| |
| 作者单位： | 内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018,内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018,内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018,内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018,内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018,内蒙古农业大学 机电工程学院/内蒙古自治区草业与养殖业智能装备工程技术研究中心, 呼和浩特 010018 |
| |
| 基金项目： | 国家自然科学基金项目（11364029）；内蒙古自治区自然科学基金项目（2017MS0606） |
| |
| 摘 要： | 为解决种公羊运动行为的识别依赖饲养员观察耗时耗力的问题,本研究设计了一种基于加速度传感器的种公羊运动行为识别系统。该系统利用无线加速度传感器节点采集种公羊的运动行为信息,对行为信息进行实时采集和无线传输,分析传感器4种部署方案下(背部、颈部、前腿、后腿)采集到的羊行为数据,并利用K均值聚类法和区间阈值分类法进行分类。试验表明传感器的4种部署方案中将传感器部署在种公羊的背部靠近前腿处得到的加速度数据最稳定。但K均值聚类法平均识别率为77.05%,识别效果差,因此又提出了区间阈值分类法,通过对加速度数据识别测试获得区间阈值,对静立、行走、奔跑行为的识别率分别达到95.96%、95.78%和96.89%,3种行为的平均识别率达到96.21%。本研究所获得的运动行为数据可应用于种公羊的运动量补充和健康状况监测。
|
| 关 键 词： | 种公羊 运动行为 加速度传感器 K均值聚类 区间阈值分类 |
| 收稿时间： | 2018/1/15 0:00:00 |
Recognition of the movement behavior of stud rams based on acceleration sensor |
| |
| ZHANG Xiyu,WU Pei,XUAN Chuanzhong,YANG Jianning,LIU Yanqiu and HAO Min.Recognition of the movement behavior of stud rams based on acceleration sensor[J].Journal of China Agricultural University,2018,23(11):104-114. |
| |
| Authors: | ZHANG Xiyu WU Pei XUAN Chuanzhong YANG Jianning LIU Yanqiu and HAO Min |
| |
| Institution: | School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China,School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China,School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China,School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China,School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China and School of Mechanical and Electrical Engineering/Inner Mongolia Engineering Research Center for Intelligent Facilities in Grass and Livestock Breeding, Inner Mongolia Agricultural University, Hohhot 010018, China |
| |
| Abstract: | In order to solve the problem recognizing stud rams'' movement behaviors depends on feeders'' observation is time consuming,a rams'' movement behavioral identification system with acceleration sensor was designed. In the system, atriaxial acceleration sensor was employed to collect the rams'' movement data in real time, and the data was wirelessly transmited. The system also could obtain the characteristics of movement behaviors by analyzing the data collected under four deployment schemes of sensor (the sensor fixed on the back, neck, foreleg and hindleg of ram), and classify the movement behaviors by K-means clustering algorithm and interval threshold algorithm. The results showed that the sensor fixed on the back near the forelegs of ram could obtain the most stable acceleration data in four deployment scenarios. The average recognition rate of K-means clustering algorithm was 77.05% and its recognition was not good. So this research presented interval threshold algorithm. The interval threshold was obtained by indentifying and testing the acceleration data. The recognition rates of standing, walking and running were 95.96%, 95.78% and 96.89%, respectively. The average recognition rate for the three movement behaviors reached 96.21%, showing that the movement behaviors data obtained by analysis could be applied to guidance in supplement of stud rams'' exercise and monitor their health condition. |
| |
| Keywords: | stud ram movement behaviors acceleration sensor K-means clustering algorithm interval threshold algorithm |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《中国农业大学学报》浏览原始摘要信息 |
| 点击此处可从《中国农业大学学报》下载免费的PDF全文 |
|
