DEV Community

Cover image for Building an AI Drone System for Real-Time Crowd Risk Detection (Edge AI + IoT)
Prithwis Das
Prithwis Das

Posted on

Building an AI Drone System for Real-Time Crowd Risk Detection (Edge AI + IoT)

We Built an AI Drone System to Detect Crowd Danger in Real Time

Crowd disasters often happen when dense groups are compressed by surrounding movement — and by the time humans notice, it’s already too late.

So we built NaeonAIr, an AI-powered drone monitoring system designed to analyze crowd risk from the sky.

It combines:

  • Computer vision (YOLO) for aerial person detection
  • Crowd flow & pressure analysis to detect dangerous buildup
  • A Jetson Nano edge gateway that can coordinate up to 100 devices
  • Smart city IoT integration (oneM2M Mobius)
  • A TypeScript web dashboard with real-time 3D risk visualization

The goal is simple:

Turn drones into intelligent safety sensors that help prevent crowd accidents before they happen.

We designed it as a scalable AIoT architecture, not just a single-drone demo — meaning it can support multi-device monitoring across large events or urban areas.

You can check out the full system breakdown, architecture, and hardware here 👇
🔗 https://www.hackster.io/choi-hyong-chan-kang-naeon-das-prithwis/naeonair-ai-eye-for-the-sky-380123

Would love feedback from the AI, IoT, and drone communities!

Top comments (0)