DroneX: Drone Detection System

Led a team of 3 electrical engineering students in designing a live video-based drone detection system with a smart recording feature.

Trained a custom YoloV5 model on a diverse collection of multiple datasets, totaling 200,000 images, achieving a real-time detection accuracy of 90% precision and 89% recall, and deployed on Jetson Nano.

Orchestrated server setup using Docker and NGINX to manage real-time video stream conversions.

Developed a responsive website to display an HLS live feed with real-time detections and historical records of previous system detections using NGINX. Link: https://www.youtube.com/watch?v=aYokR-FGdaU