AI-based Animal Health and Welfare Inspection for Export Compliance
The Australian red meat and livestock industry employs over 430,000 people and contributes more than $72.5 billion to our economy. Australia is a world leader in health welfare standards for red meat and livestock exports. Our ability to supply livestock and top-quality red meat products largely hinges on the regulatory environment, such as the Australian Standards for the Export of Livestock and The Export Meat Program which provides inspection, verification, and certification services to the industry. However, our industry is facing challenges to maintain its global competitive position based on the trusted “Brand Australia”.
What are the challenges?
- Manual livestock inspection – this is laborious, time-consuming and subjective, leading to data quality variation and discrepancies, high regulatory cost and administrative burden for processors and exporters. Since consumers are increasingly demanding ethical meat production, there is an increased need for livestock monitoring, data collection, health and welfare reporting. However, there is limited adoption of digital and emerging data-driven technologies in the industry. For example, prior to export or processing, daily health and welfare checks for all cattle are conducted by government accredited stockpersons and/or veterinarians to identify that are not fit for travel due to injury or illness. The number of cattle to be inspected for each consignment can vary from hundreds to tens of thousands. Onsite inspection of a large number of crowded cattle in harsh environments is a challenging and risky task, reliant on subjective observations to identify stock unfit for export or processing. This sometimes leads to discrepancies between the government and industry in the assessment of animal health and welfare. The cost of getting incorrect assessment, i.e., deeming the animal is or isn’t fit for processing becomes exponentially expensive.
- Growing shortage of rural vets – most graduating vets choose to work with small animals (pets) in metro areas leading to a rural vet shortage.
This project aims to employ advanced Computer Vision and Artificial Intelligence technologies to develop an automated and integrated prototype system for real-time livestock monitoring using multiple cameras. The functions of the system include establishing normal or baseline levels of certain herd behaviours, providing alerts and reporting of animal health and welfare, which can be used by farmers, trained stock inspectors, livestock vets, meat processing companies, and/or livestock exporters for export compliance.
This project will contribute to modernising Australia’s regulatory approach, reduce regulatory cost for processors and exporters, make better use of innovative technologies for robust, real-time, risk-based regulation to maintain and strengthen Australia’s global reputation as a provider of assured, high-quality meat and livestock. It is expected that the proposed system will deliver a significant benefit for the Australian red meat and livestock industry.
Our Industry Collaborators
Northern Co-operative Meat Company Ltd (CASINO Food Co-op)
Work Package 1: Data collection and annotation, led by Dr Caroline Lee
Work Package 2: Data analysis, led by Dr Dadong Wang
Work Package 3: System design and testing, led by Stuart Anderson
Project Governance: Aaron Ingham (Lead), Caroline Lee, Greg Bishop-Hurley