The Deployment of Machine Learning and On Board Vision Systems for an Unmanned Aerial Sprayer for Pesticides

In the Smart Farming (SF) domain, integrating autonomous systems is revolutionizing the efficiency and sustainability of Crop Management (CM) practices. This paper introduces an approach to Pest Control (PC) in Tea Plantations (TP), focusing on using an autonomous Unmanned Aerial Vehicle (UAV) equipped with a Pest Detection (PD) and precision spraying system. Leveraging the capabilities of the DJI Agras T40, a UAV specifically engineered for agricultural use, this system incorporates a Deep Learning (DL) built on the DenseNet-121 architecture. This model is refined to accurately detect and accurately evaluate the infection rates of six prevalent tea pests. In order to intelligently identify pesticide dispersion, the UAV uses advanced technology. This provides targeted deployment, optimizes the utilization of resources, and minimizes impact on the environment. The method's effectiveness has been proved by simulation experiments, recommending that it has real-world possibilities. A sustainable and flexible approach to several pest cases can be achieved by pairing the Sprayer Control Module (SCM) with the PD. Such integration significantly advances autonomous Pest Control Systems (PCS), enhances PC precision and performance, and minimizes the environmental impact.