|
|
The Italian start-up company Oròbix Life from Bergamo has developed an AI (artificial intelligence) system that recognises the quality of grapes in harvest boxes and thus automates the selection. The assessment is made by a digital eye that photographs the harvest boxes from above and forwards the data to a system that has previously been trained with deep learning technology. Deep Learning is a special subfield of machine learning and imitates human learning behaviour with large amounts of data. Deep Learning uses artificially generated neurons to recognise patterns and solve concrete problems from them. Against the backdrop of an increasing shortage of labour, especially during the harvest season, the method represents a serious alternative, according to Oròbix Life.
Alessandro Chiarini, Business Developer at Oròbix Life, explains: "The application is a 'smart farming' project that uses the images of an image processing system to detect low-quality grapes. The system makes it possible to objectively assess grape quality and also allows us to reconstruct the decision-making and classification process."
The first winery to test the technology is the Lunelli Group from Trentino for the production of its renowned sparkling wine Ferrari Trento DOC. Last autumn, the project was awarded the Smau innovation prize at Milan's trade fair for business innovation.
(ru / Source: Civiltà del bere; Photo: ru)