Pilot Case 1
Weed Identification and Spot Spraying
Location: Commercial farms/cooperatives in Greece
Overview
Pilot Case 1 focuses on deploying AI-enhanced robotics for accurate weed detection and removal, following the principles of precision agriculture. The proposed solution utilizes real-time AI-based weed recognition and targeted spraying, carried out by the A-200 Husky UGV equipped with a spot sprayer. By identifying and spraying only the weeds, this approach aims to drastically reduce herbicide usage, minimize environmental impact such as soil contamination, and improve the overall resource use efficiency. The integration of augmented reality (AR) further strengthens the system by giving operators a real-time, data-enriched view of the field, overlaying key information directly onto their line of sight to support better decision-making and more precise spraying.
Timeline
The Pilot Case 1 begins with farmer engagement and data collection, focusing on gathering images and on-farm data to train AI models capable of distinguishing weeds from maize crops. This is followed by algorithm refinement, during which the trained models are tested and fine-tuned in a controlled environment. Next, the technology is integrated with Automated Robotic Applicators (ARAs) to evaluate its performance in the field and iteratively improve it based on real-world results and farmer feedback. The final phase involves large-scale demonstrations to showcase the effectiveness and scalability of the solution.
Validation
The validation and evaluation of Pilot Case 1 will focus on testing the weed identification and spot spraying systems in terms of accuracy, operational efficiency, and environmental impact. Field trials will be conducted to benchmark the system’s performance against conventional methods, using metrics such as efficiency, crop protection product (CPP) usage, and effects on crop health. Farmer feedback will play a key role in ensuring the technology meets practical needs and supports broader sustainability objectives.
AR Integration
Pilot Case 1 will use augmented reality (AR) for real-time weed identification, overlaying critical data onto the user’s field of view to guide precise spot spraying by robotic systems. Additionally, AR-based simulations will be used to train users in recognizing various weed species and practicing the timely operation and configuration of the spraying robots.
