Some details are only shown to signed-in consortium members. Sign in to see them.
*Short summary of your UC* UAVWEED will publish an open, very high-resolution UAV image dataset with expert-validated weed plant annotations in winter wheat crops. The dataset builds on field campaigns, image annotation workflows and weed mapping results from weed-AI-seek and BETTER-WEEDS, and is designed to train and benchmark AI models for low-altitude UAV-based weed detection. Introduction Weeds can reduce crop yield, but they also contribute to biodiversity and ecosystem services. Practical weed management therefore needs reliable information on where weed species occur and which species require action. Today, site-specific control is limited by the lack of large, well-documented, species-level annotated image datasets for robust AI model development. Methodologies Low-altitude RGB UAV imaging in winter wheat crops from very low flight heights (1-3 m altitude) Expert-supported bounding box annotations using CVAT annotation software Quality control through field inventories, expert knowledge and georeferenced ground truth weed assessments. Integration of associated evidence from weed-AI-seek: YOLO-based real-time UAV detection and edge- computing workflows Output Open Access Annotated Weed Image Dataset with high-resolution UAV images and species-level annotations. Technical documentation and annotation guidelines covering data generation, formats, metadata and quality assurance. Impact The dataset supports AI-based weed detection for example for spot spraying, selective weeding and biodiversity-friendly crop protection. It contributes to FAIR agricultural data, improved model generalisation, reduced pesticide use, and European data-space activities linked to the Green Deal, Farm-to-Fork, CAP eco- schemes and biodiversity monitoring. What I can offer Curated UAV image data and annotation experience for weed detection in arable crops. Domain expertise in UAV field campaigns, weed mapping, ground truthing and AI model validation. What I need Support in metadata harmonisation, repository integration and long-term visibility. Exchange with partners on common annotation standards, data governance, and new EU projects.
Source: Annual AgData Use Case Summit 2026 booklet.
| Germany |
| Research Institute |
| P2 | Leibniz Institute for Agricultural Engineering and Bioeconomy (ATB Potsdam) | Germany | Research Institute |