This research project conducted by Agriculture NSW in collaboration with the University of Technology, Sydney (UTS) investigates a proof of concept for using RGBD cameras to estimate fat and muscle score in cattle. This project has built on earlier work by Agriculture NSW, using laser beam technology, to estimate hip height within the phenotypic prediction program of the Beef CRC.
Estimating fat thickness, either by hand or using ultrasound equipment requires considerable experience, with accurate operators not always readily available. Hence, a method using 3D cameras to accurately estimate fatness, integrated into a real time system, would greatly enhance the use of this vital parameter to increase productivity for the beef industry. In addition, muscle score (muscularity) is an important variable (trait) for the estimation of retail beef yield and, as a new input into BeefSpecs calculator.
Thus, this project investigates the fundamental data acquisition, image and point cloud processing techniques and formulates the problem in a machine learning space to estimate P8 Fat (mm), rib fat (mm) and muscle score (category) from cattle using RGBD cameras, without the need for ultrasonic measurements or trained assessors.
- Malcolm McPhee
- Dr Alen Alempijevic
- Brad Walmsley
- John Wilkins
- Hutton Oddy
- Brian Kinghorn
- David Mayer
- Michael Beer
- Meat & Livestock Australia (MLA)
- Australian Provisional Patent Application No. 2014903163 – “3D IMAGING” filed 26 August 2014.