Identification of naked lamb racks using machine learning algorithms

AMPC partnered with MEQ Probe to develop a suite of algorithms, software, and machine learning models to identify naked lamb rack primals on a red meat processing plant conveyor belt at chain speed. 

As part of the project, 30,000 images of naked lamb racks were collected and used for model training and analysis. Once these images were captured and labelled, a machine learning model was developed, and an algorithm was trialed to recognise naked lamb rack primals in images. 

The machine learning model was used on a boning belt with 99.9 per cent accuracy at chain speed.