Lean meat yield estimation using a prototype 3D imaging approach.

Area of Science:

• Agricultural Science
• Computer Vision
• Animal Science

Background:

• Lean Meat Yield (LMY) is a critical trait in the beef industry but is not routinely measured.
• Objective on-line technology for LMY determination is needed for wider industry adoption.
• Current methods for LMY assessment are labor-intensive and not suitable for real-time abattoir application.

Purpose of the Study:

• To present a proof-of-concept for estimating beef carcass LMY using 3D imaging.
• To develop and present an on-line data acquisition system for abattoir applications.
• To evaluate the accuracy and viability of 3D imaging for LMY estimation.

Main Methods:

• A novel on-line data acquisition system with three RGB-D cameras was designed for 3D carcass side reconstruction.
• Hindquarter segmentation was performed on 3D models to extract curvature information.
• Lean Meat Yield was estimated using non-linear regression with Gaussian Process models, incorporating curvature, P8 fat, and hot-standard-carcass-weight (HSCW).

Main Results:

• The study evaluated 119 beef carcasses across two commercial abattoirs.
• The 3D imaging approach achieved a Root Mean Square Error (RMSE) of 3.91% and an R-squared (R²) of 0.69.
• Results indicate a viable and relatively accurate technology for estimating LMY.

Conclusions:

• 3D imaging using RGB-D cameras provides a promising on-line solution for estimating Lean Meat Yield in beef carcasses.
• The developed data acquisition system and estimation model demonstrate the feasibility of implementing this technology in commercial abattoirs.
• This approach can lead to more efficient and accurate LMY assessment, benefiting the Australian beef industry.