Weed Detection and Removal System for Smart Agriculture Using Deep Learning

The rise of artificial intelligence (AI) and Internet of Things (IoT) technologies is transforming agriculture by enabling smarter, more efficient farming practices.This paper presents the development of an autonomous robotic system designed for precision weed detection and selective herbicide application in agricultural fields.The proposed system uses a robotic vehicle equipped with a deep learning-based vision system to detect weeds and a selective spraying mechanism that targets only unwanted plants, thereby reducing herbicide usage and minimizing environmental impact.The robotic vehicle navigates autonomously within the field using waypoint navigation, capturing real-time video through a mounted camera, which is processed to identify and locate weeds.Leveraging AI-based inference results, the selective spraying system activates only on detected weeds, enhancing resource efficiency and reducing manual labour.This approach provides a scalable and sustainable alternative to traditional weed management practices, supporting farmers in reducing costs and improving field productivity.Field testing demonstrates the effectiveness of the system in autonomously navigating varied agricultural environments and accurately performing selective weed spraying.