The AgriGaia platform enables accessible, cloud-based AI development for agriculture by offering an open-source data space for dataset management, model training, and deployment. Built on GAIA-X principles of interoperability and data sovereignty, it supports collaborative, modular, and reproducible AI workflows. This paper demonstrates its use in agricultural robotics for detecting Rumex weeds in pastures, a task typically requiring manual labour. A mobile robot equipped with a camera and GPU runs two YOLO-based models trained on locally stored data within AgriGaia: one for general object detection, and another specialized for Rumex identification. Containerized training ensures reproducibility, while deployment through Portainer allows flexible switching and live monitoring of models. This setup illustrates how AgriGaia lowers technical barriers, promotes collaboration across research and industry, and empowers farmers to apply adaptable AI solutions that enhance efficiency and sustainability in precision agriculture.
Gerwin et al. (Thu,) studied this question.