Synthesis of an algorithm for determining the coordinates of the center point of the udder and top of cow’s teats in a robotic milking installation
Автор: Shilin Denis Viktorovich, Vasiliev Alexey Nikolaevich
Журнал: Вестник аграрной науки Дона @don-agrarian-science
Рубрика: Электротехнологии, электрооборудование и энергоснабжение агропромышленного комплекса
Статья в выпуске: 2 (66), 2024 года.
Бесплатный доступ
The machine vision system in robotic milking machines plays a crucial role in the automation of the milking process, and is usually built on the basis of technologies such as RGB cameras, infrared thermal imaging cameras, ToF cameras and RGBD cameras. It allows the robot to accurately determine the location of the cow’s udder and control the milking process. The machine vision system reduces the risk of errors and improves the quality of milking, as well as the health and welfare of animals. Thus, the machine vision system helps make the milking process more efficient, safe and comfortable for both animals and farmers. Improving methods for detecting the middle of the udder, necessary for localizing the teats, and the coordinates of the top of the teats makes it possible to increase the speed and reduce the accident rate of the milking robot in unfavorable conditions. A current and promising direction in the development of robotic milking technology is to improve these algorithms, increase their operating accuracy and fault tolerance. The results of the operation of the main modules of the algorithm for detecting the top of the teats during robotic milking based on polynomial interpolation and singular value decompositions are presented. The work demonstrates the clear advantage of synthesized algorithms over analogues under unfavorable conditions on the farm, which affect the quality of recording a machine vision system. The proposed algorithms can be used as the basis for an intelligent robotic milking system as feedback to control the position of the manipulator’s working body. The use of such a system in field conditions in combination with sensors for continuous monitoring of quantitative and qualitative indicators of milking will significantly increase work productivity and reduce the number of idle passes in the milking installation.
Milking installation, robotic milking, machine vision, computer vision, point cloud segmentation, polynomial interpolation, singular value decompositions
Короткий адрес: https://sciup.org/140305992
IDR: 140305992 | DOI: 10.55618/20756704_2024_17_2_66-76