SkySAR

Inspiration

Thousands of people go missing every year in forests and national parks. Current search methods are slow, risky, and limited by human capacity. We wanted to create a faster, scalable, and safer solution.

What it does

SkySAR is an autonomous drone swarm system that divides search areas into grids and scans them in parallel. Equipped with AI, LiDAR, and cameras, it detects missing people and relevant objects while reducing human risk and search time.

How we built it

We developed a prototype using Gemini Vision Pro for multi-class visual detection and a Streamlit dashboard for real-time visualization. The system simulates drone coverage and alerts, demonstrating AI-assisted decision support.

Challenges we ran into

Ensuring accurate detection in dense forests and variable weather

Coordinating multiple drones and search grids in simulation

Balancing data processing needs with limited bandwidth

Accomplishments that we're proud of

Built a functional prototype demonstrating AI-powered search and detection

Created a real-time dashboard for operators

Developed a scalable concept that could integrate hundreds of drones in future deployments

What we learned

Effective collaboration is key when integrating AI, hardware, and UX design

Ethical and privacy considerations are critical in drone surveillance

Simulating real-world SAR operations helps refine AI models and workflows

What's next for SkySAR

Integrate autonomous drone swarms with on-board AI

Optimize AI models for real-time human detection in forests

Expand testing in controlled environments and eventually pilot deployments with SAR teams

Built With

• gemini
• python
• streamlit