UAV Based Farm Inspection using Deep Learning

Recent advancements in AI and UAV (Unmanned Aerial Vehicle) technology have enabled real-time precision farming applications, transforming traditional agricultural practices. This paper presents a novel farm inspection system combining UAVs, deep learning, and AI edge devices to analyze crops and livestock in real time. Utilizing the lightweight YOLOv8n object detection model on the NVIDIA Jetson Orin Nano, the system achieves rapid detection of crop diseases, plant health issues, and livestock conditions within 90 ms and an overall accuracy of $\sim 93 \%$. The UAV provides aerial data for large-scale monitoring, while the edge device processes data locally, reducing latency and enabling autonomous operation in remote areas without cloud connectivity. This integration enhances inspection speed, coverage, and accuracy, enabling timely interventions and resource optimization. By leveraging deep learning, edge computing, and UAV technology, the proposed system demonstrates significant potential to improve agricultural efficiency and sustainability.