When artificial intelligence lends winegrowers a helping hand

he head of the R&D department at Sun’Agri, Damien Fumey, explains that “artificial intelligence helps us to create the ideal micro-climate for vines”. For the past 15 years, the company has collected billions of data items to develop models that are not only “energy-giving and based on solar tracking” but truly “agronomic”, providing finely-tuned shading or a reduction in solar panels depending on the requirements of vine growth and the weather.

“As frost approaches, the blinds automatically switch to flat towards the end of the night to increase the greenhouse effect and maintain a +2°C gap in the protected vine buds”, explains Fumey. The same process protects vines from heat spikes. “The gaps can range from up to -10°C at canopy level to -30°C at soil surface level”, he adds.

“By equipping harvesting machines and hoppers used at Domaine du Chapitre with low-cost GNSS tracking boxes, and by geo-locating the reception areas, we used algorithms to successfully link the information provided by weighing tickets to vine blocks”, explains Romain Girardot, research engineer with Montpellier SupAgro, detailing findings from the Celectiv programme conducted in conjunction with the French Vine & Wine Institute (IFV) to create granular yield mapping at inter-block level. The resultant mapping allows the vintage to be characterised, technical choices to be compared and agronomic decisions to be adjusted the following season.

Currently, 90% of data generated by the wine industry is neither structured nor usable, and less than 1% of data is analysed, despite the fact that industry players need advances in technology to be able to cope with challenges like climate change and a reduction in inputs. Other applications are in the pipeline, including detecting grapevine trunk diseases and flavescence dorée along with learning how to prune vines.