ForestGuard is an open-source, full-stack satellite monitoring and environmental change detection platform. It couples high-resolution optical imagery (Sentinel-2) and spectral index analysis with an interactive geospatial dashboard, change vectorization, automated timeline tracking, and AI-assisted environmental impact reporting.

• Interactive Geospatial AOI Selection: Draw arbitrary polygon boundaries directly on an interactive MapLibre GL map or load pre-configured benchmark sites (Amazon Rainforest, Elephant Butte Reservoir, Austin Urban Growth).
• Multi-Temporal Spectral Analysis: Computes pixel-level difference masks across core spectral indices:
  • NDVI (Normalized Difference Vegetation Index) for canopy disturbance and deforestation.
  • MNDWI / NDWI (Normalized Difference Water Index) for drought, reservoir shrinkage, and flood recession.
  • NDBI (Normalized Difference Built-up Index) for urbanization and bare-ground expansion.
• Vectorized Hotspots & GeoJSON Overlay: Converts pixel change masks into GeoJSON polygons with computed geographic coordinates, centroid locations, and hectare-scale area measurements.
• Interactive Split Slider: Real-time before/after satellite composite inspection with interactive wipe-slider comparisons.
• Chronological Change Timeline: Visualizes historical NDVI progression, event severity, and state transitions over the selected observation window.
• Grounded AI Synthesis: Generates structured executive summaries detailing key findings, critical zones, environmental impact, and verification steps based strictly on calculated metrics.
• Executive Report Generation: Exports standalone, publication-ready HTML environmental audit reports containing executive summaries, KPI breakdown tables, change catalogs, and verification recommendations.
• Resilient Fallback Mode: Gracefully transitions to offline benchmark datasets when backend services or satellite APIs are unreachable.

forestguard/
├── src/                     # Frontend (Next.js 16 + React 19 + MapLibre GL)
│   ├── app/                 # App Router (Landing, /monitor, /cases)
│   ├── components/          # Analysis controls, map viewers, split-screens, reports
│   └── lib/                 # API client, domain adapters, context, mocks
└── backend/                 # Backend (FastAPI + Pydantic v2 + Geospatial Pipeline)
    ├── app/
    │   ├── api/             # REST endpoints (health, analyses, reports)
    │   ├── services/        # Spectral indices, change detection, vectorization, AI
    │   └── schemas/         # Pydantic DTO models
    └── tests/               # Pytest suite for API & change detection algorithms

• Framework: Next.js (App Router, Turbopack)
• Styling: Tailwind CSS
• Maps: MapLibre GL
• Icons: Lucide React

• Framework: FastAPI (Python 3.10+)
• Validation: Pydantic v2
• Geospatial Processing: NumPy, SciPy (ndimage.label), Shapely
• Satellite Provider: Google Earth Engine (Sentinel-2 SR Harmonized) with built-in offline synthetic provider (MockSatelliteProvider)

• Node.js: v18.17+ or v20+
• Python: v3.10+
• npm or pnpm / yarn

1. Navigate to the backend directory:

   cd backend

2. Create and activate a Python virtual environment:

   # Linux / macOS
   python -m venv .venv
   source .venv/bin/activate
   
   # Windows (PowerShell)
   python -m venv .venv
   .venv\Scripts\Activate.ps1

3. Install backend dependencies:

   pip install -e .

4. Configure environment variables (optional for local testing; defaults to mock provider):

   # Windows PowerShell:
   Copy-Item .env.example .env
   # Linux / macOS:
   cp .env.example .env

5. Start the FastAPI development server:

   uvicorn app.main:app --reload --port 8000

   Backend is available at http://localhost:8000 (Interactive Swagger docs: http://localhost:8000/docs)

1. Open a new terminal in the repository root:

   # Install dependencies
   npm install

2. (Optional) Verify frontend environment variable in .env.local:

   NEXT_PUBLIC_API_URL=http://localhost:8000

3. Start the Next.js development server:

   npm run dev

   Frontend is available at http://localhost:3000

• Open http://localhost:3000 and click "Launch Monitor" or navigate directly to /monitor.

• Quick Benchmark: Click one of the benchmark site buttons:
  • 🌲 Amazon (Deforestation in Para, Brazil)
  • 💧 Reservoir (Elephant Butte water loss, NM)
  • 🏗️ Urban (Suburban expansion in Austin, TX)
• Custom Area: Click "Draw Polygon on Map" and click on the map canvas to construct a custom boundary. Double-click or click the first point to close the polygon.

• Select the Baseline (Before) observation date window (e.g. 2025-03-01 to 2025-03-31).
• Select the Target Observation (After) date window (e.g. 2026-08-01 to 2026-08-31).

• Toggle desired environmental change layers:
  • 🌲 Forest Loss (NDVI anomaly)
  • 💧 Water Loss (MNDWI reduction)
  • 🏗️ Built / Bare (NDBI expansion)
  • 🛣️ Road / Mining (SWIR linear features)

• Click "Run Environmental Analysis".
• Watch the live step-by-step pipeline progress:
  1. Acquiring satellite imagery
  2. Removing cloud contamination
  3. Generating spectral indices
  4. Comparing temporal imagery
  5. Classifying environmental changes
  6. Mapping change regions

• Change Hotspots: Click individual polygons on the map or in the change catalog to zoom to the hotspot, inspect change magnitude, and view spectral delta scores.
• Before / After Slider: Toggle the Split-Screen Comparison tab to drag the interactive wipe-slider over satellite composites.
• Observation Timeline: Inspect temporal NDVI trajectories and status milestones.
• AI Summary: Review auto-generated impact assessments, critical vulnerability zones, and validation notes.
• Export Audit Report: Click "Generate Report" to produce an executive HTML report with summary stats and change tables.

cd backend
python -m pytest tests/ -v

npm run build

Distributed under the MIT License. See LICENSE for details.