WildGuard Random Forest Fire Detection Model
This repository contains a trained RandomForestClassifier used by the WildGuard system to detect high-probability wildfire events from NASA FIRMS satellite data.
Model Details
- Algorithm: Random Forest
- Features:
- latitude, longitude
- bright_ti4, bright_ti5
- frp, scan, track
- daynight, hour, month
- Output: Probability of wildfire event
Intended Use
- Real-time wildfire monitoring
- Alerting systems
- Research & academic projects
Limitations
- Dependent on satellite resolution and FIRMS latency
- Not a replacement for official emergency systems
Associated Project
Frontend & backend available in:
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support