Procedures for measuring 3-D surface area and surface temperature of livestock
Bio-energetic models of livestock require geometrical properties (e.g., surface area) of animals. Currently, the surface area of animals is estimated from empirical equations based on body weight or assuming the animal as cylinders or spheres. These approximations do not accurately characterize the true geometry of the animal. Similarly, skin-surface temperature predictions from bio-energetic models are validated by taking spot measurements, which neglects the spatial variability of this parameter. In this study, an accurate approach of measuring these parameters is proposed. Based on preliminary results (more data analysis is ongoing), the advantages and disadvantages of various 3-D scanning technologies (photogrammetry, Kinect V1, Kinect V2, Intel® RealSense™ SR300, and Intel® RealSense™ R200) as well as two different procedures for obtaining 3-D surface temperature (thermal projection and thermal photogrammetry) are discussed. Kinect (V1 or V2) is recommended for 3-D scanning of livestock because the sensors are cheap and easy to use. Sense from 3-D Systems (Intel® RealSense™ SR300) does not support large scanning volumes but it is recommended for scanning small animals (or parts of an animal) because it is handheld and the software is easy to use. To scan large animals (e.g., a whole cow) and for advanced users, Intel® RealSense™ R200 is suitable. For 3-D surface temperature, thermal photogrammetry is recommended because it is less error prone than thermal projection.