Fraunhofer Center for Assistive Information and Communication Solutions – AICOS
01.09.2026
At Fraunhofer Portugal AICOS, we believe that cutting-edge technology holds the key to preserving and elevating traditional industries. Through the VineKeeper project, our research team is leading a technological shift in Portuguese viticulture, combining world-class artificial intelligence with practical digital tools. Co-funded by the European Union, this 36-month initiative running from January 2026 to December 2028 is transforming how native grapevine varieties are identified, cataloged, and protected.
Traditional ampelography and phytosanitary diagnostics rely heavily on manual, time-consuming field observations by rare specialists, while alternative genetic testing remains costly and slow. Fraunhofer Portugal AICOS is tackling these bottlenecks head-on by leading the scientific research and development of advanced Computer Vision and Artificial Intelligence algorithms.
Our work goes beyond theoretical models. Leveraging state-of-the-art architectures like Vision Transformers (ViTs) and Domain Adaptation techniques, our research—backed by foundational studies such as the ViTiFound thesis—has successfully achieved exceptional cross-border generalization across field and laboratory datasets. Our core scientific and technological contributions include:
> Automatic Grapevine Identification Algorithms: Engineering models capable of recognizing roughly 200 native Portuguese grape varieties directly in real-world field conditions.
> Disease Detection Modules: Building robust classifiers to distinguish healthy from non-healthy plants and target specific pathogens impacting vegetative propagation.
> OIV Knowledge Base Integration: Studying the application of Vision-Language Models (VLMs) to extract standardized descriptors from the International Organisation of Vine and Wine (OIV).
> Smart Acquisition Support: Developing mobile-side algorithms that guide users during data collection, automatically evaluating image quality and focusing on key vegetative structures like leaves, clusters, and tendrils.
"The VineKeeper project aims to contribute to the preservation, enhancement, and sustainability of Portugal's winemaking heritage by providing winemakers with advanced digital tools for the automatic identification of native grape varieties and the early detection of diseases," highlights Ivo Façoco, Researcher at FhP-AICOS.
Strategic Collaboration with GEODOURO and the Consortium
A cornerstone of FhP-AICOS’ mission is ensuring that high-impact research seamlessly translates into market-ready solutions. Through a close technological partnership with GEODOURO—the project promoter responsible for platform development and commercial exploration—the advanced capabilities developed by our researchers are being directly integrated into the SIGP (Sistema Integrado de Gestão de Propriedades) mobile and web platform.
This powerful synergy bridges the gap between advanced research and agricultural end-users, supported by the viticultural expertise of Plantivet and validated in real-world operational environments by esteemed strategic stakeholders, including Quinta do Crasto, Esporão S.A., and Italy’s Vivai Cooperativi Rauscedo (VCR).
"We expect this solution to reduce the use of phytopharmaceutical products, support the conservation of minority or endangered varieties, and boost the overall efficiency and competitiveness of wine production. VineKeeper stands out for its integrated use of Artificial Intelligence and Computer Vision to automatically identify native varieties and vineyard diseases from field-captured images. Its integration with the SIGP platform and its focus on preserving Portugal's wine heritage reinforce its truly innovative character.", explains the researcher.
By combining Fraunhofer Portugal AICOS’ scientific rigor with GEODOURO’s robust platform architecture, VineKeeper is democratizing technical expertise, cutting operational costs, and reinforcing Portugal's position at the forefront of digital agriculture and precision viticulture.