Robotic Weed Removal Using Deep Learning for Precision Farming

This research explores the integration of intelligent robotics and deep learning for precise weed management in cotton field. Utilizing the YOLOv5 model, the study evaluates real-time weed detection, weed removal efficiency and achieving high accuracy across varied weed types in controlled environments. Key contributions include the development of a delta robot system for automated weed removal and the utilization of the “CottonWeeds” dataset, enabling robust model training. Findings underscore the importance of accurate positional error testing and calibration for effective robotic deployment. The study advocates for sustainable agricultural practices by minimizing herbicide use through targeted weed control strategies. Future research aims to enhance model accuracy and scalability, advancing the field of precision agriculture towards increased crop productivity and environmental sustainability.