Naked primal cut recognition vision system trial in plant

25 September 2020
Focus area: Advanced manufacturing
Program stream: Hands-off processing
Project number: 2018-1048
AMPC partnered with Strategic Engineering to trial a naked primal cut recognition vision system in an Australian red meat processing plant. The system used 3D imaging sensors to capture and process dimensions, fat content and shape to identify the naked primal cut. 

As part of the project, 1169 images of primal cuts were gathered on site by the vision system. Using these images, machine learning software was trained to identify seven different primal cuts.

The results from the trial found that the recognition software was able to identify the primal cuts with an accuracy of 90 per cent. 

Supporting documents

Most recent reports

Previous in this focus area 29 October 2020 First prototype automation for deboning lamb shoulder stage 2 Next in this focus area 18 November 2019 Automation of primal cut bagging