IgnisGuard AI

Inspiration

Wildfires can spread rapidly while traditional detection and emergency dispatch systems remain fragmented and slow. We envisioned an AI system that could continuously monitor camera feeds, satellite hotspots, and weather conditions, then transform those signals into actionable emergency intelligence within seconds.

What it does

IgnisGuard AI provides real-time wildfire detection, fire-spread forecasting, and automated emergency dispatch. It combines lookout camera analysis, NASA FIRMS satellite hotspots, and NOAA weather data to detect smoke, estimate fire behavior, identify threatened areas, and generate CAP alerts, WEA messages, and ICS-209 tactical briefings.

How we built it

We built IgnisGuard AI with React, TypeScript, Express.js, Google Gemini 3.7 Flash, NASA FIRMS, NOAA weather data, BigQuery, Firestore, and Google Cloud Run. Gemini handles multimodal camera analysis, structured incident reasoning, multilingual alerts, and tactical assistance, while backend services process telemetry and generate standardized emergency outputs.

Challenges we ran into

The biggest challenges were distinguishing smoke from fog and atmospheric inversions, processing high-resolution video with low latency, grounding AI decisions in real telemetry, and generating alerts that satisfy strict emergency communication formats and character limits.

Accomplishments that we're proud of

We demonstrated a sub-30-second detection-to-alert pipeline, generating broadcast-ready CAP 1.2 alerts in under 15 seconds. We also implemented multilingual warnings, real-time incident visualization, tactical AI assistance, and an emergency operations dashboard.

What we learned

We learned how multimodal AI can combine visual, satellite, weather, and geospatial data for real-time disaster intelligence. We also gained practical experience integrating AI with standardized emergency protocols and building low-latency, cloud-based public-safety systems.

What's next for IgnisGuard AI

Next, we plan to integrate certified emergency alert gateways, autonomous drone verification, edge AI deployment on lookout towers, and advanced wildfire perimeter simulations using real-time terrain, wind, and fire-behavior data.

Built With

• bigquery
• computer-vision
• docker
• express.js
• fema-ipaws
• firestore
• gemini-3.7-flash
• google-cloud-run
• google-genai-sdk
• ics-209
• latex-math
• leaflet.js
• lucide-react
• multimodal-ai
• nasa-firms-api
• noaa-raws-weather
• oasis-cap-1.2
• react-19
• tailwind-css
• typescript
• vite
• wea-sms
• web-audio-api
• web-speech-api