Introduction of ReWeeMap project AI driven remote sensing for Datura Stramonium detecting

Abstract

The ReWeeMap project, co-financed by EIT Food and the European Union, aims to develop an AI-driven, image recognition-based software capable of specifically detecting Datura stramonium (jimsonweed) among selected arable crops using supervised learning. To achieve this, we employed advanced drone technology equipped with multispectral imaging cameras.Field testing and data collection were conducted at two locations in Hungary, focusing on maize, tomato, and pepper crops. Meanwhile, the software development and IT implementation were carried out in Lithuania by BetaVia (formerly ART21). The development was scientificly proved by the University of Szeged. Following nearly one year of data collection and development works, the software now operates with an accuracy exceeding 80%.The project’s long-term objective is continuous improvement, a core principle in AI-based software development. The methodology established for this software is scalable and adaptable to detect other hazardous weed species in various crop types. Consequently, the project envisions not only the commercialization of the software but also its expansion based on the developed framework.Ultimately, this initiative contributes to reducing the food industry’s exposure to contamination by toxic weeds by enabling early-stage detection and mitigation directly at the source of infestation. Our aim is to present the development steps and methodology we have established, with the hope that it will support the broader adoption of digital solutions designed to mitigate threats to the agri-food sector.