Research/Using "machine vision" to count hens and reduce egg breakage - proof of concept
From Poultry Hub
Title: Using "machine vision" to count hens and reduce egg breakage - proof of concept
Project Leader: Greg Cronin at AWSC funded by Poultry CRC (Project 06-11)
Duration of Project: Now completed (1 July 2006 to 31 August 2007)
Contents |
Project overview
Machine vision is defined as the ability of a computer to “see”. In a machine vision system, video cameras supply information to a computer, and depending on the software, objects can be recognised, tracked and measured. For the modern cage egg industry, machine vision offers a means to increase the level of surveillance of the birds and mechanised processes such as egg collection, without the need to increase unit labour inputs.
Project objectives
This project aims to test the proof-of-concept that machine vision in the modern cage egg industry can be used to:
- Count live laying hens in cages; and
- Identify egg belt blockage.
Project progress
This project has been completed and the Final Report is available from the Poultry CRC (ISBN: 1 921010 05 3).
This one-year proof-of-concept project began in July 2006. The project developed a prototype machine vision system to recognise targets on live hens in cages, enabling the computer to count the number of live hens in the cage, and to recognise potential blockages on the egg collection belt. The prototype is being tested under commercial conditions.
Related news
Using machines to monitor hen welfare and reduce egg breakage Posted on 24th January by Chris Day, Poultry CRC
Results from a recently completed Poultry CRC proof-of-concept project, Using "machine vision" to count hens and reduce egg breakage, has demonstrated that ‘machine vision’, or video image analysis (VIA – the ability of a computer to ‘see’), can be used to monitor egg belts for potential blockages and has the potential for effectively counting hens.
Testing under commercial conditions has demonstrated that the ability of a machine to automatically scan the egg belt to identify foreign (non-egg) objects has improved to a 95% success rate.
The Project Leader, Greg Cronin, from DPI Victoria’s Animal Welfare Science Centre (AWSC), says that the results indicate that VIA can be successfully used to monitor activities within the modern cage-layer shed.
“Continued refinement of the software and hardware will further improve the detection rate,” said Greg, “with potential benefits measured not just in terms of economics, but in more varied and interesting jobs for stockpeople, which may also result in better retention of staff.”
As Australia’s population, and its workforce, ages, the pool of personnel available to the poultry industries is expected to shrink, and this technology may be part of a solution to the problem.
“VIA technology could also be applied to automatically monitor other production processes in the poultry and other livestock industries,” said Greg, “providing benefits from improved frequency of surveillance and enabling more time to be made available to the stockperson for less-mundane tasks. We consider counting hens in cages a first step in being able to monitor hens automatically, potentially leading to a mechanism for automatic monitoring of hen welfare and health.”
While VIA has already been used to grade eggs after collection, the use of VIA to reduce the incidence of cracked or broken eggs prior to collection does not appear to have been considered before this project.
For more information, contact Greg on (03) 9742 0450 or email Greg.Cronin@dpi.vic.gov.au
