Recognition of Cow Teats Using the 3D-ToF Camera When Milking in the “Herringbone” Milking Parlor

The foreign robotic milking systems are presented in the Russia’s agricultural market. These systems have a high cost and low capacity relative to the “Herringbone” milking parlor. It is necessary to search for new solutions for robotic milking parlor of “Herringbone” type, based on the manipulator and vision system. This article presents a method of processing a cloud of points obtained from a 3D ToF camera installed in the technical vision module. By applying the k-nearest neighbors method and the “Euclid distance” method, it was performed the segmentation of the points cloud in order to detect the teats of the cow’s udder.