UWE Bristol’s Centre for Machine Vision team and other members of a consortium has been awarded Innovate UK funding to develop a low cost, augmented reality picking aid that will display information about berry maturity through the use of machine learning and spectral imaging cameras.
The consortium includes AR developers Opposable Games; environment, food and science research organisation NIAB EMR and leading industry grower-owned co-operative, Berry Gardens Growers Limited, alongside the Centre for Machine Vision which is part of the Bristol Robotics Laboratory.
The concept and commercial opportunity was identified by Richard Harnden, Director of Research at Berry Gardens Growers Ltd who has wanted to improve the consistency of the eating quality of the co-operative’s premium berry lines, which includes a sweet eating dessert blackberry, for several years.
“It is very hard for pickers, especially new pickers, to really understand the correct stage of ripeness in the blackberry before picking it”, he said. “Pick it too early and, although the berry will be black in colour, it won’t have accumulated enough sugars and so it will still taste acidic. Pick it too late, and the berry will be too soft to withstand the supply chain and will leak juice in the punnet.”
He continued, “There is a small correct window for picking the fruit that delivers an exquisite combination of sweetness and flavour, which can be done by eye but it takes time for pickers to achieve the correct level of perception. The proposed picking aid, using novel technology, will deliver a maturity indicator, which will guide new and experienced pickers alike to quickly make the right decision every time.”
Bo Li, a machine vision specialist in the Centre for Machine Vision at Bristol Robotics Laboratory at UWE Bristol, who devised the project, said: “By developing a low cost multispectral camera for detecting the real time ripeness of fruit, we can enhance the efficiency of picking, reduce the requirement for pickers to be experienced, and shorten the training time required. This step forward will improve the consistency of fruit quality and customer satisfaction.”
In an industry already experiencing difficulties in accessing experienced staff, the impact of Covid-19 is putting additional strains on farms and farm workers. Restrictions on labour movement, new safety measures, and risk mitigation procedures being required, mean that the horticultural and agricultural industries must look to novel solutions to train new workers and meet existing and future labour requirements. Global demand for high quality and healthy food such as soft fruit is increasing. To meet this demand farms are looking to technological solutions that enable increasing the quality, yields, and productivity whilst reducing environmental impacts. This project will contribute towards the UK government’s Transforming Food Production objectives, part of the Industry Strategy Challenge Fund.
The Innovate UK funded project will commence in September 2020, with the development of a prototype device building on the experience of the consortium, then moving on to field trials. Members of Berry Garden Growers Ltd will trial the harvesting aid on their farms as the project progresses.
The Centre for Machine Vision is part of the Bristol Robotics Laboratory (BRL). We solve real-world practical computer vision problems. Their particular excellence lies in three-dimensional reconstruction and surface inspection. They are recognised as one of only three UK centres with expertise in Photometric Stereo (PS). They have pioneered PS in industry, medicine and defence/security. Their laboratory supports REF (Research Excellence Framework) level research activities and research-led teaching in machine vision. Find out more here.