Eagle Eye: Drone With Advance AI Integration

The integration of artificial intelligence (AI) with unmanned aerial vehicles (UAVs) has significantly enhanced drone autonomy and real-time decision-making. However, many existing systems rely heavily on cloud computing, are cost-prohibitive, or lack flexible manual control. This paper introduces Eagle Eye, a manually piloted, edge-AI-powered drone system designed for real-time vision-based surveillance and multipurpose field applications. Built on a Raspberry Pi 5 (12 GB RAM) paired with a Google Coral USB Accelerator and a Sony high-definition webcam, Eagle Eye executes efficient on-board computer vision models provided by the AI Makers Kit, enabling tasks such as object detection, tracking, and scene understanding.The aerial platform integrates a SpeedyBee F7 V4 flight controller, a SpeedyBee 50A 4-in-1 ESC, DYS G-power series brushless motors, 10-inch Orange HD propellers, and is powered by a custom-designed 6-cell 22.2V Li-ion battery with a 5V buck converter supporting auxiliary electronics. The chassis features a lightweight fiberglass landing gear for structural resilience. Manual flight control is maintained using a FlySky FS-i6X transmitter, while AI-enhanced video feeds are transmitted via a VNC-based remote desktop application running on 64-bit Raspberry Pi OS.Designed for accessibility and affordability, Eagle Eye achieves real-time inference exceeding 30 FPS at 720p resolution, maintaining an end-to-end processing latency of under 100 milliseconds. Maintaining a human-in-the-loop architecture, the system empowers operators with AI-augmented visual feedback for quick decision-making without relying on external cloud infrastructure. Field-testing demonstrates Eagle Eye’s suitability for disaster response, agricultural monitoring, and high-security surveillance, proving that advanced AI functionality can be effectively deployed on compact aerial platforms without compromising performance, reliability, or costefficiency.