Machine learning for beef production efficiency (PhD)

Improving the efficiency of beef production through the use of precision farming technology and machine learning techniques

The Challenge

In 2018, 43% of prime beef cattle sent to slaughter in the United Kingdom (UK) did not meet target specifications (AHDB, 2019), with 11% being too fat and 32% achieving poor conformation grades. This shows that almost half of the UK’s prime beef cattle are being sent to slaughter by their producer despite the fact they will not achieve target carcass grades. These over-/under-finished cattle result in negative financial implications along with increased greenhouse gas emissions.

Under-finished cattle tend to have a lower saleable meat yield percentage, meaning the cuts are smaller and of a lower quality, resulting in a lower UK demand and therefore lower prices.  Over-finished cattle are on farm longer than necessary, meaning increased inputs such as feed, bedding and labour, which is estimated to cost UK producers £8.8 million per year (AHDB, 2018). Over-finished cattle also result in more greenhouse gas emissions per head per lifetime, with a 14-day over-finished animal resulting in 230kg CO₂ eq per head (SRUC Data).

The Project

Due to these inefficiencies, farmers aim to achieve the target carcass specifications that will result in premium grades and therefore premium prices. However, the opening figure shows this is not successfully being achieved. This is likely due to current on-farm assessment methods not performing accurately, resulting in poor production efficiency. Therefore, new assessment methods must be explored such as the use of precision livestock farming (PLF) technology and machine learning techniques to create prediction algorithms. This will hopefully provide farmers with an accurate and objective means of determining when their cattle are ready for slaughter and will have the potential to replace current subjective methods. This will also allow farmers to accurately achieve the target grades, reducing the number of over-/under-finished cattle and therefore improving profitability while reducing greenhouse gas emissions.

Student

Holly Nisbet, SRUC