From detection to action: Creating operational prescription maps for weed management using low-cost UAVs

Weeds remain one of the leading constraints to sustainable agriculture, prompting widespread herbicide use that poses environmental and economic concerns. Precision agriculture, particularly Unmanned Aerial Vehicle (UAV)-based remote sensing, offers a promising alternative by enabling site-specific weed management. However, despite their cost-effectiveness relative to ground-based systems and their growing technical feasibility, adoption rates in Italy remain below 10 %. This limited uptake stems from complex interrelated barriers: fragmented land ownership, generational divides, lack of awareness and trust in emerging technologies. Cultural resistance and perceived technical complexity further hinder integration at the farm level, especially among smallholders. This study investigates the operational potential of low-cost UAVs combined with open-source image classification algorithms to produce prescription maps for targeted herbicide application. Specifically, it aims to assess the feasibility of using commercially available, low-cost drones equipped with standard cameras, together with GIS tools, to reduce herbicide use and improve environmental sustainability. The effectiveness of this approach is evaluated across multiple spatial configurations and weed infestation thresholds, with a view toward broader applicability in Italian and comparable agro-environmental contexts. The experimental site is set on a maize field in NE Italy, the research tested three classification methods—Artificial Neural Networks (ANN), Maximum Likelihood Classifier (MLC), and Object-Based Image Analysis (OBIA)—across varying grid sizes and weed detection thresholds to provide modular prescription maps for different weed management scenarios. Results demonstrate a significant reduction in the treated area compared to conventional blanket spraying, with ANN ranging from 18.92 % to 3.75 %, MLC from 14.59 % to 2.18 %, and OBIA from 15.82 % to 3.49 %, depending on the configuration. This study highlights how commercial low-cost UAVs and open-source GIS tools, when applied through a structured and reproducible workflow, can bridge the gap between innovation and practical application in sustainable agriculture