Real-Time Beef Bone Belt Monitoring Solution: Red to White Analysis (Phase 1)

This project aims to develop and trial a real-time feedback concept for fabrication room supervisors to monitor the efficacy of beef boners responsible for shoulder blade deboning and ultimately improve yield recovery during the beef boning process. Maintaining consistent high-quality deboning practices among staff, even when supervisors are not present, is a challenge. This project will evaluate the ability of an off-the-shelf AI camera to provide real-time feedback on potential yield loss on a bone belt in a beef fabrication environment, beginning with the deboning of shoulder blade bones (this project) and subsequently extending to hind leg bones (future project). If successful, the project could result in a 20% improvement at the JBS Brooklyn facility site, with an estimated annual benefit of $174,096 for shoulder blade bones and $539,136 for forequarter and hindquarter legs combined. JBS has been developing a range of applications for AI camera technologies, supported by Cognex. The Cognex development team and JBS continue to refine the onboard AI software for the meat industry, a sector previously not focused on by Cognex. This has led to an AI/ML camera strategy supported by Cognex engineers. This project represents the first phase of a multi-phase design program, starting with the procurement of a lab-based Cognex camera and the development of AI-trained algorithms to evaluate its suitability and effectiveness for the Australian red meat processing industry. Successful validation from this project will provide key learnings for future development.