Measurement of 3-D surface area of piglets

Abstract

Precision livestock farming is focused on identifying and monitoring animals individually to improve comfort and production of each animal. One of the key parameters that influence animal comfort and production is heat stress. In the U.S., heat stress costs the swine industry $300 million per year. One way to study alternatives for reducing heat stress in swine is studying their heat balance using bioenergetic models. Bioenergetic modeling of animals requires surface area and geometrical properties of the animals being modeled. The current practice for obtaining surface area of animals is through empirical equations developed more than 50 years ago. The validity of these equations on application of present-day animals, which have changed because of genetics and nutrition, is questionable. Other approach has been to assume the animal geometry as a cylinder or sphere. The objective of this study is to determine the surface area of piglets using images from 3-D cameras and to compare the results against surface areas obtained using the current practice. The 3-D geometry of animals was obtained using Kinect V1 camera (Xbox 360) with ReconstructMe software and ZED stereovision camera. Post-processing of 3-D geometric meshes was performed in MeshLab and Blender. The preliminary data measured using 3-D scans vs. current practices agreed, on average, within 1 cm for length, 5 cm for circumference, 12 dm2 for surface area, and 6 dm3 for volume.