🚨 RockShield AI

AI-Powered Rockfall Prediction & Real-Time Alert System for Open-Pit Mines

Overview
Features
Recent Updates
Architecture
Tech Stack
Project Structure
Quick Start
Usage Guide
Screenshots
API Documentation
Contributing
License
Team & Contributors
FAQ

Overview

RockShield AI is a comprehensive rockfall prediction and real-time alert system designed specifically for open-pit mining operations. The platform combines AI/ML models, IoT sensor integration, satellite imagery, and interactive dashboards to help mining operations predict rockfall hazards, analyze slope stability, and maximize worker safety.

Real-time risk assessment, interactive satellite mapping, detailed analytics, and automated alerts - all in one scalable, low-cost, life-saving solution.

Features

🎯 Core Capabilities

🤖 AI-Powered Risk Prediction
- Machine learning models (XGBoost, Random Forest) for rockfall probability assessment
- Slope stability analysis with Factor of Safety (FoS) calculations
- Real-time risk classification (High/Medium/Low)
- Confidence scoring and probability metrics
🛰️ Interactive Satellite Map
- Live satellite imagery using ESRI World Imagery
- Terrain view toggle for topographical analysis
- 9 coal mining regions across India with risk zone overlays
- 285+ active mine locations tracked
- Color-coded risk zones (red = high risk, green = medium risk)
- Clickable zones/markers with detailed risk information
- Interactive popups showing zone names, mine counts, and risk scores
📊 Risk Analytics Dashboard
- Comprehensive analytics with 6 interactive charts
- Risk distribution across Indian mining locations
- Monthly incident trends and seasonal patterns
- Risk trigger analysis (rainfall, earthquake, human activity, natural erosion)
- Location-based filtering and statistics
- Bar charts, pie charts, area charts, and line graphs
🔮 Real-Time Prediction System
- Rockfall risk assessment with customizable parameters
- Slope stability calculations with Factor of Safety (FoS)
- Circular gauge visualizations for instant risk assessment
- Color-coded results (red/orange/green)
- Professional empty states and loading animations
- Contextual safety recommendations
🌐 Modern Web Interface
- Beautiful gradient backgrounds with glassmorphism effects
- Smooth Framer Motion animations throughout
- Responsive design for desktop, tablet, and mobile
- Consistent design language across all pages
- Intuitive navigation with quick action buttons
- Professional typography using Poppins font family

📡 Alert & Communication

Automated SMS/Email Alerts (Twilio/SendGrid integration ready)
Risk threshold notifications
Real-time alert triggers based on ML predictions

📊 Analytics & Insights

Explainable AI with SHAP and LIME integration
Historical trend analysis
Risk pattern recognition
Safety margin calculations

Recent Updates

✨ Latest Features (November 2025)

🛰️ Satellite Map Feature

Added interactive satellite map with live imagery
Integrated 9 coal mining regions across India
Implemented risk zone overlays with clickable polygons
Added satellite/terrain view toggle
Created GeoJSON data structure for 285+ mine locations
Built "Insights to Safety" educational panel

📊 Risk Analytics Dashboard

Comprehensive analytics page with 6 interactive charts
Location-based filtering system
Monthly trends and seasonal analysis
Risk trigger breakdown (rainfall, earthquake, human activity, erosion)
Statistics cards showing risk distribution
Real-time data visualization

🎨 Prediction Page Interface Redesign

Complete UI overhaul with glassmorphism effects
Replaced helicopter icon with mountain emoji (⛰️)
Circular gauge visualizations for results
Improved form inputs with icons
Enhanced loading states and animations
Professional empty states

Architecture

System Components

Frontend Layer (React + TypeScript)
- Hero page with feature highlights
- Interactive satellite map
- Risk analytics dashboard
- Prediction interface
- Navigation and routing
Backend Layer (Python Flask)
- REST API endpoints
- ML model serving
- Data processing pipeline
- Analytics calculations
ML Models (XGBoost, Random Forest)
- Rockfall risk classification
- Slope stability analysis
- Feature engineering
- Model inference
Data Layer
- GeoJSON risk zone data
- Mining location database
- Historical incident data
- Sensor data integration (ready)

Data Flow

User Input → Frontend → REST API → ML Models → Predictions → Frontend Display
                                          ↓
                                   Analytics Engine
                                          ↓
                                   Risk Assessment
                                          ↓
                            Alert System (if threshold exceeded)

Tech Stack

Frontend

Framework: React 18 with TypeScript
Build Tool: Vite
Styling: TailwindCSS
Animations: Framer Motion
Maps: Leaflet.js + React-Leaflet
Charts: Recharts
Icons: Lucide React
Routing: React Router DOM

Backend

Framework: Flask (Python)
ML Libraries: scikit-learn, XGBoost, pandas, numpy
Model Persistence: joblib
API: RESTful endpoints

DevOps & Tools

Linting: ESLint
CSS Processing: PostCSS
Package Manager: npm

Integrations (Ready)

SMS Alerts: Twilio
Email Alerts: SendGrid
Explainable AI: SHAP, LIME
Satellite Data: ESRI World Imagery, ISRO Bhuvan (configurable)

Project Structure

RockShield-AI/
├── src/                          # Frontend source code
│   ├── Components/              # React components
│   │   ├── Hero.tsx             # Homepage hero section
│   │   ├── SatelliteMap.tsx     # Interactive satellite map
│   │   ├── RiskAnalytics.tsx    # Analytics dashboard
│   │   ├── PredictionPage.tsx   # Risk prediction interface
│   │   ├── Navigation.tsx       # Navigation bar
│   │   ├── Footer.tsx           # Footer component
│   │   └── ...
│   ├── data/                    # Static data and GeoJSON
│   │   └── coalRiskZones.ts     # Mining regions data
│   ├── Types/                   # TypeScript type definitions
│   ├── App.tsx                  # Main app component
│   ├── main.tsx                 # Entry point with routing
│   └── index.css                # Global styles
│
├── ml-service/                  # Python ML backend
│   ├── app.py                   # Flask REST API
│   ├── models/                  # Trained ML models
│   ├── generate_models.py       # Model training script
│   ├── requirements.txt         # Python dependencies
│   └── templates/               # HTML templates (if needed)
│
├── models/terranox/             # Terranox ML integration
├── tests/                       # Test files
├── .config/                     # Configuration files
├── package.json                 # Node dependencies
├── tsconfig.json                # TypeScript configuration
├── vite.config.ts               # Vite configuration
├── tailwind.config.js           # Tailwind CSS config
├── .env.example                 # Environment variables template
├── README.md                    # This file
├── IMPLEMENTATION_SUMMARY.md    # Satellite map documentation
├── RISK_ANALYTICS_FEATURE.md    # Analytics documentation
├── PREDICTION_PAGE_UPDATE.md    # Prediction page documentation
└── RockShield-AI.png            # Project screenshot

Quick Start

Prerequisites

Node.js (v18 or higher)
Python (v3.8 or higher)
npm or yarn
pip

Installation

1. Clone the Repository

git clone https://github.com/Adesh2204/RockShield-AI.git
cd RockShield-AI

2. Set Up Frontend

# Install dependencies
npm install

# Start development server
npm run dev

The frontend will run on http://localhost:5175/

3. Set Up ML Backend

# Navigate to ml-service directory
cd ml-service

# Install Python dependencies
pip install -r requirements.txt

# Start Flask server
python app.py

The backend API will run on http://localhost:5001/

4. Access the Application

Homepage: http://localhost:5175/
Satellite Map: http://localhost:5175/satellite-map
Risk Analytics: http://localhost:5175/analytics
Prediction: http://localhost:5175/predict

Usage Guide

Navigating the Platform

Homepage

View feature highlights and system overview
Access all major features via action buttons
"Launch Live Demo" → Go to prediction page
"Risk Analytics" → View analytics dashboard
"Satellite Map" → Explore interactive map

Satellite Map (`/satellite-map`)

View Risk Zones: Red polygons indicate high-risk areas, green for medium risk
Toggle Views: Switch between satellite and terrain imagery
Click Zones: Get detailed information about mining regions
Explore Locations: 9 regions across India with 285+ mine locations
Navigate: Use quick action buttons to access other features

Risk Analytics Dashboard (`/analytics`)

View Statistics: High risk %, medium risk %, regions monitored
Filter by Location: Dropdown to view specific region data
Analyze Charts:
- Bar chart: Risk distribution by location
- Pie chart: Overall risk breakdown
- Area chart: Monthly incident trends
- Line chart: Risk score progression
- Horizontal bar: Risk trigger analysis
Hover for Details: Interactive tooltips on all charts

Prediction Page (`/predict`)

Rockfall Risk Assessment:

Fill in parameters:
- Latitude and Longitude
- Rainfall (mm)
- Triggers (count)
- Landslide size
- Administrative division
Click "Predict Risk Level"
View results:
- Risk percentage (circular gauge)
- Classification (High/Medium/Low)
- Probability and confidence metrics

Slope Stability Analysis:

Fill in technical parameters:
- Cohesion
- Friction angle
- Unit weight
- Slope height and angle
- Water pressure
- Reinforcement type
Click "Calculate Factor of Safety"
View results:
- Factor of Safety (FoS) value
- Stability status (Unstable/Marginal/Stable)
- Safety margin assessment
- Interpretation guide

API Endpoints

Rockfall Risk Prediction

POST /predict
Content-Type: application/json

{
  "latitude": 22.5726,
  "longitude": 88.3639,
  "rainfall": 150.5,
  "triggers": 3,
  "landslide_size": "medium",
  "administrative_division": "West Bengal"
}

Response:
{
  "risk_probability": 65.4,
  "risk_class": "MEDIUM",
  "confidence": 0.87
}

Slope Stability Calculation

POST /predict-fos
Content-Type: application/json

{
  "cohesion": 25.0,
  "friction_angle": 30.0,
  "unit_weight": 18.5,
  "slope_height": 15.0,
  "slope_angle": 45.0,
  "water_pressure": 50.0,
  "reinforcement": "anchors"
}

Response:
{
  "factor_of_safety": 1.45,
  "status": "MARGINAL",
  "safety_margin": "Acceptable with monitoring"
}

Analytics Data

GET /analytics

Response:
{
  "location_data": [...],
  "monthly_trends": [...],
  "risk_distribution": [...],
  "triggers": [...],
  "summary": {...}
}

Screenshots

Interactive dashboard showing DEM-based mine map with AI-driven risk heatmap and real-time notifications.

API Documentation

For detailed API documentation, see:

IMPLEMENTATION_SUMMARY.md - Satellite map feature
RISK_ANALYTICS_FEATURE.md - Analytics dashboard
PREDICTION_PAGE_UPDATE.md - Prediction interface

Contributing

Contributions are welcome! To contribute:

Fork the repository
Create a feature branch:
```
git checkout -b feature/amazing-feature
```
Commit your changes:
```
git commit -m "Add amazing feature"
```
Push to the branch:
```
git push origin feature/amazing-feature
```
Open a Pull Request

Development Guidelines

Follow TypeScript/React best practices
Maintain consistent code style (ESLint)
Write descriptive commit messages
Test thoroughly before submitting
Update documentation as needed

License

This project is licensed under the MIT License.
See LICENSE file for details.

Team & Contributors

Core Team

Adesh - @Adesh2204 - Lead Developer

Special Thanks

To all contributors and supporters who made this project possible!

FAQ

Q: What sensors are compatible with RockShield-AI?
A: Any device capable of sending data in JSON format. Mock scripts for Arduino/Raspberry Pi are provided for testing. Real sensor integration requires basic API connectivity.

Q: How do I deploy this in production?
A:

Frontend: npm run build and serve the dist/ folder
Backend: Use Gunicorn/WSGI with Flask, set up proper environment variables
Database: Configure production database for sensor data
Security: Implement authentication, HTTPS, rate limiting

Q: Can I customize the ML models?
A: Yes! Train new models using ml-service/generate_models.py or replace existing models in ml-service/models/. The system supports any scikit-learn compatible model.

Q: How accurate are the predictions?
A: The models are trained on synthetic data for demonstration. For production use, retrain with real-world data from your specific mining operations for best accuracy.

Q: Can I add more mining regions?
A: Yes! Edit src/data/coalRiskZones.ts to add new GeoJSON polygons and markers for additional regions.

Q: Is there a mobile app?
A: Currently web-only, but the responsive design works well on mobile browsers. A dedicated mobile app is planned for future releases.

Q: How do I integrate real IoT sensors?
A: Configure sensor devices to POST data to the Flask API endpoints. The .env.example file shows required environment variables. Contact the team for sensor integration guides.

Q: Can I use this for other types of mining?
A: Yes! The system is adaptable. You'll need to retrain models with relevant data and update the risk zone definitions for your specific mining type.

Contact & Support

For questions, issues, or feature requests:

GitHub Issues: RockShield-AI/issues
Email: Contact via GitHub profiles

Built with ❤️ for safer mining operations worldwide

Last Updated: November 6, 2025

FilesExpand file tree

README.md

Latest commit

History