P.PSH.1418 - Nuctech DEXA CT system validation for beef chining automation

Summary

To support future beef boning development, it is proposed that Nuctech will carry out a series of testing using a CT machine to investigate and validate the ability of the device to generate high quality and accurate data for the automation system. Nuctech's DEXA CT machine has been selected to trial for the conceptual design and to examine the machine's capabilities. The main goal of this series of trials is to obtain enough information, validate assumptions, and gain confidence for the investment decision. The outcome of the phased testing will be the demonstration of application of CT to investigate and validate the ability of the device to generate high quality and accurate data for the automation system.

Objectives

The objective of this project was to assess the image quality of the Nuctech CT system installed in an Australian abattoir using the test devices available within the XTE-CT test piece, including the image resolution, dimension estimation, stability across the scan field and repeatability. This test piece has been adopted as the calibrating device for CT scanners used to provide reference data for technologies seeking commercial accreditation for predicting carcase lean and fat% in Australia. The ability to differentiate fat, lean and bone tissue and thereby to determine CT composition %, bone landmarks and muscle seams in Nuctech CT scans of beef primals was also assessed, in addition to the quality of imaging beef sides scanned in five sections.

Key findings

1. Good image resolution and the ability to measure dimensions in a scaled and highly repeatable fashion, as demonstrated through scans of the XTE-CT test piece, demonstrates the capacity of the Nuctech CT system to provide imagery suitable for automation.

2. Fat, lean and bone tissues were well differentiated using pixel value thresholds in Nuctech CT scans of bone-in short loin and rib set primals. The Nuctech CT scanner was thus able to predict the medical CT composition (fat, lean and bone %) of these beef primals with excellent precision.

3. The ability to differentiate tissues in Nuctech CT images of beef rib sets scanned continuously over a 5 minute period was maintained, while differentiation of fat and lean tissue was slightly reduced in the largest sections when entire beef sides were scanned in five sections.

Benefits to industry

The consistent ability of the Nuctech CT to differentiate bone from soft tissue in scans of all thicknesses in this study is very promising for use of this imaging system to drive automated beef boning therefore potentially unlocking this step change for industry.

MLA action

MLA is working with AMPC to align the industries beef boning strategy.

Future research

Together with this it is recommended that methods to integrate this sensing system with robotic cutting modules be investigated.