Weight estimation system of Japanese black cattle using non-contact computer vision

已经开发了使用计算机视觉技术的牛的测量系统。在日本黑牛的生长过程中，定期对其品质进行评价是非常重要的。不仅检查牛的体重和大小，而且检查姿势和形状，作为主要评价标准。本研究利用三维计算机视觉设备从不同角度同时获取牛的数字点云数据。这些装置预先校准以具有共同的坐标系。将从不同角度获得的点云数据在一个公共坐标系上进行重构，并考虑牛的方位在计算机中生成数字牛。在计算机上重建了牛的模型，估计了评价牛的姿势的重要指标。所获得的数据将是估计牛体重的重要参数。

A measurement system for cattle using a computer vision technique has been developed. It is important to evaluate the quality of Japanese black cattle periodically during the growth process. Not only the weight and size of cattle, but also the posture and shape, are checked as the primary evaluation criteria. In this study, the digital point cloud data of cattle are obtained by three-dimensional computer vision devices from different angles simultaneously. These devices are calibrated in advance to have a common coordinate system. The point cloud data obtained from different angles is reconstructed on a common coordinate system, and digital cattle are generated in the computer considering the orientation of the cattle. The important specifications for evaluating the cattle posture are estimated from the reconstructed cattle model on the computer. The obtained data will be an important parameter for estimate the weight of the cattle.