AI camera technology enhances integrity and market access

In a major step forward for supply chain transparency and product integrity, Meat & Livestock Australia (MLA) and the Australian Meat Processor Corporation (AMPC) have partnered with Retail Ready Operations Australia (Coles RROA) to trial a new off-the-shelf deep learning smart camera system.

The Cognex In-Sight D900 – a smart camera embedded with ViDi deep learning software – utilises innovative technology designed to perform complex in-line inspection tasks.

Moreover, it has been trained for red meat application to improve label accuracy and identify meat cuts and primals in beef and lamb processing, helping safeguard the reputation of Australian red meat domestically and internationally.

A smarter system for a complex task

Traditional machine vision systems used in meat processing rely on rule-based algorithms and are typically limited to label inspection.

These systems often struggle with environmental variations such as lighting and orientation,and lack the flexibility to be trained for more complex tasks like identifying specific meat cuts.

The newly trialled AI-powered smart camera system overcomes these limitations. It integrates deep learning capabilities into a streamlined hardware and software package, making it significantly easier to implement in red meat processing environments.

The system demonstrated exceptional performance during the trial, achieving a false positive rate of less than 0.29%, while maintaining a 100% true positive and 0% false negative rate – an impressive leap forward in accuracy and reliability.

Protecting market access and brand reputation

Incorrect labelling and product errors remain a persistent challenge for the Australian red meat industry.

Domestically, these issues can lead to costly product recalls and damage to brand reputation.

Internationally, the stakes are even higher as labelling errors can result in the loss of export licenses and reduced access to key markets.

With approximately 70% of Australian beef and 78% of lamb destined for export, maintaining product integrity is critical.

Accurate labelling not only ensures compliance with market requirements but also reinforces Australia’s global reputation for quality, safety, and trustworthiness.

By investing in this technology, MLA and AMPC are helping the industry protect its competitive edge and uphold its status as a premium protein supplier.

This has been highlighted with the camera’s ability to recognise Chinese characters for export label inspections.

Next steps for industry adoption

Following the successful trial, MLA and AMPC will continue collaboration to support broader adoption of smart camera systems across the red meat industry.

This system’s ease of use and rapid training capabilities make it well-suited for integration into diverse processing environments.

Future efforts focusing on tailoring technologies such as these to meet the specific needs of individual supply chains are encouraged – ensuring processors and brand owners across the sector can benefit from their capabilities.

This initiative marks a significant advancement in the use of AI for red meat processing and highlights the industry’s commitment to innovation, digitised integrity, and global competitiveness.