Analyze → Predict
Multi-Modal AI Risk Engine
Every signal is normalized into an explainable contribution. Risk score changes as environmental inputs change.
Multi-modal AI assessment
Rockfall risk score
out of 100 · North Wall B
Detected reasons
- All monitored signals are within the operating safety envelope
Fusion pipeline
A transparent simulation of the ROCKGUARD decision path.
Camera
Weather
Temperature
IoT mesh
Context
AI model output: 28/100 · Safe
The model favors immediate exclusion when camera fracture confidence, rain loading, and geotechnical movement rise together.
Risk Factor Contribution
Every factor contributing to the current score
Camera vision
8% fracture / rock-change confidence
1.6 pts
Weather
1.2 mm/h rain · Clear
0.8 pts
Thermal change
2.1°C change over 3 hours
2.6 pts
IoT sensors
3.5 mm displacement · 2.4 mm/s vibration
10.5 pts
External context
12 mm / 72h history · geology score 32
6.1 pts
Risk classification rules
Operational bands used for this prototype.
Safe
0–34Continue monitored operations.
Warning
35–59Increase inspection frequency.
High
60–79Establish controlled exclusion zone.
Critical
80–100Alert workers and evacuate immediately.