基于光流场的奶牛监控识别系统设计

DESIGN OF COW MONITORING AND RECOGNITION SYSTEM BASED ON OPTICAL FLOW FIELD

为了解决现有奶牛个体信息接触式获取移动性差、费时费力等缺点,搭建了基于MATLAB的奶牛视频监控硬件系统,并借助于GUI(图形用户界面)开了奶牛视频监控界面,实现了对奶牛无接触、无干预监控。基于运动目标检测的改进光流场算法,实现了对视频中奶牛运动目标的监测、提取和识别,为视频中奶牛的个体识别、高级行为感知与分析奠定了基础。经实际运行表明开发的系统实时性好,运行可靠稳定,监控效果理想,能够实现对奶牛的实时远程监控。

In order to solve the disadvantages of poor mobility,time-consuming and laborious in the individual information contact acquisition of cows,a video monitoring hardware system based on MATLAB was built,and the video surveillance interface was opened with the help of GUI（ graphical user interface） to realize the non-contact and nonintervention monitoring of cows. The monitoring,extraction and recognition of the moving target of the cow in the video were realized based on improved optical flow algorithm of moving object detection,which laid the foundation for the individual identification and advanced behavior perception and analysis of cows in video. The actual operation showed that the developed system had good real-time performance,reliable and stable operation,ideal monitoring effect,and can realize real-time remote monitoring of cows.