Weed detection and herbicide recommendation using YOLO

In modern agriculture, efficient weed management is critical to maximizing crop yield and minimizing environmental impact. Traditional methods of herbicide application often involve uniform spraying across fields, which leads to overuse of chemicals, increased costs, and ecological harm. To address this, our research proposes an intelligent weed detection system that utilizes computer vision and machine learning to identify and count weeds in agricultural fields. The motivation behind this study stems from the existing gap in precision agriculture where real-time, automated weed quantification is limited, and current systems lack adaptability to diverse field conditions. Our proposed methodology involves capturing field images using drones or mobile platforms, preprocessing these images to enhance clarity, and employing a convolutional neural network (CNN)-based model to system's high accuracy in weed detection, achieving an overall classification accuracy of 97.3% thereby offering a scalable and eco- friendly solution for precision weed management.