Clustering of 3D point cloud data of Holstein species using black and white mottled pattern

-It may be possible to identify individuals of cattle by pattern-

Abstract

In recent years, many attempts have been made to automate cattle management using IT technology. To automate the management of cattle population, it is necessary to label each cow individually for captured data, but it is difficult to use the ear tags which is major method in Japan and QR code. In this study, individual identification is carried out using white-and-black patterns unique to Holstein cows, and clustering of three-dimensional point cloud data for each individual. We converted the three-dimensional point cloud data into a two-dimensional image representing only the pattern, and clustered by sequential method, resulting in high accuracy. The experimental results demonstrate the possibility of individual identification using white-and-black patterns.