UK potato processor Branston has teamed up with Ishida Europe and the University of Lincoln on a £823,277 project to enhance food safety and quality inspection for food manufacturers.
The partnership is exploring how to design more flexible and comprehensive vision software capable of quickly detecting faults in food packaging and products on the production line.
The 'Trainable Vision-Based Anomaly Detection and Diagnosis' project is part-funded by the UK's Technology Strategy Board and also supported by the Agriculture and Horticulture Development Board.
"Branston is supporting this project as it will greatly enhance the capacity of current systems to give feedback on the root-causes of differences in quality," said Vidyanath Gururajan, IT and projects director for Branston.
"We also recognise the value of monitoring bag sealing to minimise defects as poor seals cause rejection and generate waste."
Dr Tom Duckett, from Lincoln University's School of Computer Science, is leading the research to create new multi-purpose imaging technology that can easily be set up and applied to a variety of inspection tasks. He is also a reader and director of the Lincoln Centre for Autonomous Systems Research.
"Industrial machine vision applications are limited because of their specificity, so once you have set up the software for one product and process it doesn't really work for anything else," said Duckett.
"We are trying to make an application that is much easier to establish and more flexible to detect a variety of issues in food products and packaging."
'Commercialise the software'
The initiative is focusing on monitoring packaging seal integrity and applying the technology for this task in other ways, such as checking the right amount of food is in the right place, he said. "The aim is that Ishida will want to commercialise the software as part of its product range."
Ishida aims to work closely with the university to create the next generation of quality control systems for the food industry through the project, which is scheduled to run until October 2015.
The company said the system under development would also be more compact, resulting in more economic use of factory space.
"There will be increased assurance for customer health and safety through enhanced product quality and increased efficiencies in food production will lead to reduction in waste going to landfill," said Gary Tufnell of Ishida Europe.
Two previous pieces of research
The project builds on two previous pieces of research carried out by the University of Lincoln. The first resulted in a prototype computer vision system that can identify sub-standard potatoes. The second looked at using laser scatter imaging for detecting faults in the heat seals of food packaging.
Duckett added: "It's a case of bringing all the previous research together to create a versatile and flexible imaging technology, which is unavailable today. The current systems require manual calibration and have limited accuracy ...
"... This system will be more consistent and reliable. It should also lead to less food waste as manufacturers can detect problems earlier in the production cycle, which in turn will also lead to safer food and make the whole process much more efficient."