Back to R&D main

P.PSH.1302 - LEAP III - Sensing and stabilisation upgrade

Scott Automation & Robotics and Meat & Livestock Australia (MLA) have created an upgrade kit for existing and future LEAP III Primal systems for automated lamb processing.

Project start date: 03 May 2021
Project end date: 30 November 2023
Publication date: 01 May 2024
Project status: Completed
Livestock species: Lamb
Relevant regions: National, International
Download Report (3.1 MB)

Summary

This project will develop improved DEXA bone identification methods through skeletal identification software to improve accuracy and rib counting as well as upgrade the network communications software to achieve identified reliability improvements.

Accurate primal cutting in the Leap III module has a benefit in excess of $2 per carcase to the $7.70/head overall attributable to yield alone for the Scott LEAP sweep of modules. Skeletal identification can be further improved by utilising the material differentiating properties of dual energy X-ray combined with the use of artificial intelligence (AI) techniques including machine learning. These techniques have been rapidly improving in industry over recent years.

Objectives

The major objectives of this project are
1. Design and build retrofittable packages for the critical upgrades identified above for existing and future LEAP systems.
2. Demonstrate the upgrades to industry through installation of an existing LEAP system. The final report will detail upgrade packages that have been developed.

Key findings

Scott Technology Ltd and Meat & Livestock Australia (MLA) have created an upgrade kit for existing and future LEAP III Primal systems for automated lamb processing. The upgrades target improvements in design that increase the benefit from accuracy, achieve a reliable 10/min throughput, as well as improving mechanical reliability (uptime efficiency). The upgrade kit consists of three separate packages:
1. A software upgrade to use dual energy X-ray images, in combination with AI techniques to improve cut accuracy.
2. A carcase stabilization upgrade was designed, built and deployed at sites in New Zealand and Australia. The stabiliser was found to stabilise carcases more effectively through the X-ray scanning process, which is a key parameter in achieving accurate cuts.
3. An automated lubrication system was designed, built and deployed at sites in New Zealand and Australia. The indicator lights on the automated lubrication system provide a clear and obvious indicator that shows when a lubrication cannister requires repairing.
The three upgrade packages are all able to be retrofitted to existing machines or included in future machines.

Benefits to industry

The vision analysis upgrade package provides a significant direct benefit to the customer, as independently assessed by Greenleaf Enterprises Pty Ltd. Additionally, as the vision analysis upgrade package requires no hardware changes it is an upgrade that can be quickly installed at existing sites or new installations.

MLA action

MLA is currently looking at collaborating with Scott Automation & Robotics again for future research projects in the processing productivity area.

Future research

Significant improvements have been achieved in this project. Given the improvements over the current systems the carcase stabilisation and automated lubrication packages are now standard in future LEAP III primal systems and the analysis upgrade is available as an optional upgrade to provide additional benefit to the customer.

During this project several future improvements were discussed:
• Additional data labelling for the AI analysis. Current neural networks can achieve super-human labelling results if provided enough training data from a pool of manually labelled data. While significant improvements were achieved with the current data set, increasing the size and variability of the data set would likely increase the performance of the system. An automated labelling system as proposed in the point above would be one option to obtain more data but would require significant research and development while additional manual labelling is more quickly attainable in the short term.
• Centralisation of parts of the lubrication system. The current lubrication upgrade solves many of the problems of the current lubrication system. However, it still requires dozens of individual canisters spaced throughout the machine, all of which require monitoring and replacement. An investigation into a larger reservoir system for elements of the lubrication system could cut down on this complexity and allow simpler maintenance.
• Automated lubrication warnings. An important benefit of the lubrication upgrade was the warning lights which notify maintenance staff of empty canisters. This is a consequential improvement from the current system where maintenance staff must assess the levels of the clear lubricant within the clear canister. A further improvement would be outputs from the lubrication’s canisters to the PLC. This would allow the maintenance staff to get a notification from the machine HMI about which lubrication canisters require replacement.

More information

Project manager: Darryl Heidke
Contact email: reports@mla.com.au
Primary researcher: Scott Automation & Robotics Pty Ltd