基于Kinect的母猪呼吸频率测定算法

发布时间：2018-02-05 00:22

本文关键词： Kinect 母猪呼吸频率动态区域检测算法 DBSCAN密度聚类算法离散傅里叶变换　出处：《南京农业大学学报》2017年05期 　论文类型：期刊论文

【摘要】：[目的]呼吸频率是评价母猪健康的重要指标。针对目前畜牧业对母猪呼吸频率无接触测定的需求,提出一种基于Kinect的无接触测量算法。[方法]使用Kinect采集梅山母猪侧卧时的深度图像。通过动态区域检测算法确定腹部区域,采用DBSCAN密度聚类算法得到腹部区域质心进行局部深度图像提取。最后基于腹部区域局部深度图像的深度均值,使用离散傅里叶变换得到深度-时间拟合曲线,通过极值点确定1次呼吸过程的周期进而求出监测时间段内母猪的呼吸频次,最后转化成母猪的呼吸频率(RF)。[结果]将检测结果与人工计数结果做比较,呼吸频率正确率为85.3%,该呼吸频率测定算法是有效的。[结论]提出了一种呼吸频率测定方法,该方法实现无接触测量,准确率高,为定位饲养条件下的梅山母猪呼吸频率的测定提供了理论支撑。
[Abstract]:[Objective] Respiratory frequency is an important index to evaluate sows' health. Aiming at the requirement of non-contact measurement of respiratory frequency in sows in animal husbandry at present, a contactless measurement algorithm based on Kinect is proposed. [Methods] the depth images of Meishan sows were collected by Kinect and the abdominal area was determined by dynamic region detection algorithm. DBSCAN density clustering algorithm is used to get the center of mass of the abdominal region to extract the local depth image. Finally, based on the depth mean value of the local depth image of the abdominal region. The depth-time fitting curve was obtained by using discrete Fourier transform (DFT), and the respiration frequency of sows in monitoring period was calculated by determining the cycle of one respiration process by extreme point. Finally, it was transformed into the respiratory frequency of sows. [Results] compared with manual count, the accuracy of respiratory frequency was 85.33.The algorithm was effective. [Conclusion: a method of measuring respiratory frequency is proposed. The method has high accuracy and provides theoretical support for the determination of respiratory frequency of Meishan sows under the condition of location-feeding.
【作者单位】：南京农业大学工学院/江苏省智能化农业装备重点实验室;
【基金】：国家重点研发计划项目(2017YFD0701600) 江苏省农机三新工程项目(NJ2016-10) 中央高校基本科研业务费专项资金(KYGX201704)
【分类号】：S828;TP391.41
【正文快照】： 猪肉是我国居民重要的消费肉类,随着自动化、信息化技术在养殖领域中的应用,传统的养殖方式将向福利养殖、精细养殖方式转变。目前的规模化养殖条件下,呼吸系统疾病在母猪中发病率高、致死率高。在母猪的产前行为中,其分娩前呼吸速率会明显加快,因此母猪的呼吸速率是评价母猪

【相似文献】