A system for automatically trimming vegetables with high degree of accuracy is among the finalists for the Machinery Innovator category for this year’s British Farming Awards.
Paul Wilson’s of Scorpion Vision normally focuses on systems in food and factory environments. The new vegetable system is currently being used to peel brussels sprouts, top and tail swedes and trim the root end of leeks with ‘sub-millimetric accuracy’. The system uses a combination of 3D vision and deep learning to locate the cutting point on vegetables. The whole shape is measured before the machine decides about where to cut. According to Paul, deep learning using neural networks enhances the cutting decision as the system can be trained to understand the features the camera sees and make a more intelligent decision on where and how to cut.
The system came about when a business partner who specialises in machinery construction was approached by a grower who was already struggling to maintain a seasonal workforce for harvesting and trimming operations. The new system not only removes the manual labour from a skilled but uncomfortable working environment, but it improves health and safety as trimming no longer relies on manual cutting with large blades.
Such an automation system, which comprises a robot or servo arms to undertake the cut, typically costs around £200,000, while the machine vision system with 3D and deep learning is an additional £50,000 or less.
Photo source: Scorpion Vision