Corn Regional Optimized Weed Decisions (CROWD): a tool for a site-specific weed management

Weed management (WM) remains a primary challenge in contemporary agriculture, particularly within the European Union's Farm to Fork strategy, which aims to reduce pesticide usage by 50% by 2030 while maintaining high crop productivity. In order to achieve this objective, innovative and sustainable approaches are required; among these, Site-Specific Weed Management (SSWM) is regarded as a promising solution. SSWM employs precision agriculture technologies, including remote sensing and artificial intelligence, to optimise herbicide application by targeting only weed-infested areas. The methodology is comprised of three fundamental phases: i) Weed Detection (WD), ii) estimation of potential crop yield loss due to weeds, and iii) precision herbicide application using ISOBUS sprayers. Despite the strides made, the adoption of SSWM is impeded by the substantial costs associated with technology, its intricate nature, and its incompatibility with less digitised farming systems. This study proposes a cost-effective and rapid-deployment Decision Support System (DSS) for maize cultivation that requires minimal calibration. Building on a previously validated method, the system estimates Weed Green Cover (WGC) using RGB drone imagery by subtracting Maize Green Cover (MGC) from Total Green Cover (TGC). The economic intervention threshold, expressed as tolerable WGC, is used to define a Green Damage Threshold (GDT) for each MZ. This enables timely and targeted weeding interventions based on image-derived metrics, offering a scalable solution for sustainable WM in diverse agricultural contexts.