基于视觉定位的牧场奶牛自动跟踪定位系统研究

Study on automatic tracking and location system for dairy cows from a ranch based on visual positioning

机器视觉已经成为现代畜牧业中的一个重要应用技术,这主要是得益于近年来图像处理技术的快速发展以及计算机的硬件成本不断下降和其运算速度的飞速提升。随着近几年具有高科技的大型奶牛养殖基地的不断兴起,基于机器视觉的奶牛自动定位系统成为新的研究课题。为了解决在牧场中对特定奶牛的定位跟踪问题,笔者提出采用视觉机器人,而视觉定位系统是整个机器人系统的核心和关键。本研究利用高斯金字塔LK光流法实现对奶牛的目标定位及运动跟踪,对传统算法的不足之处进行了总结分析,提出了改进的Harris特征点检测算法。结果表明:改进后的算法不仅在特征点提取效果方面有显著提高,同时具有处理速度快的优点,可为移动奶牛的准确定位做好准备。

Machine vision has become an important technical application in modem livestock technology, which is mainly due to the rapid development of image processing technology, as well as the computer hardware costs continuing to decline and the rapid increase in computing speed. With the constant rise of large - scale and high - tech dairy cow breeding base, the automatic dairy cow positioning system has become a new research topic based on machine vision. To solve the positioning and tracking problem for specific dairy cows in the ranch, the author proposed the use of visual robot, while the vision positioning system is the core and key to the whole robot system. The study achieved the target positioning and motion tracking to the dairy cows by Gaussian pyramid LK optical flow method. The inadequacies of traditional algorithms were summarized and analyzed in this study, and then an improved Harris feature point detection algorithm was proposed. The results showed that the improved algorithm not only had a significant increase in aspect of the extraction effect of feature point, but also has the advantage of fast processing speed, which can be ready for accurate positioning of moving dairy cows.