☀️ GeoAI Solar System: Solar Potential Intelligence Engine
Version 6.0 Ultimate – Powered by SAM 3 Semantic Segmentation with advanced solar analytics including yield estimation, payback period, and carbon offset.
💀 SYSTEM OVERVIEW
GeoAI Solar System is a high‑performance geospatial intelligence platform that uses Meta’s Segment Anything Model 3 (SAM 3) to detect, segment, and analyze rooftop solar panels and potential installation zones from satellite, aerial, and UAV imagery. It transforms raw pixels into actionable business metrics: usable area, panel count, annual yield, payback period, and carbon offset.
⚡ CORE CAPABILITIES
- Pure SAM 3 Semantic Segmentation – Text‑prompt concept detection (
"rooftop solar panels", "building rooftops") with pixel‑perfect masks.
- Sliding‑window tiling – Ensures high resolution for large areas.
- Thermal health classification – Simulates hot‑spot detection using brightness (Magma palette) and maps to professional fault categories.
- Obstruction Mapping – Detects chimneys, HVAC units, and trees that create “keep-out” zones.
- Usable Area Calculation – Applies fire safety setbacks to compute net installable area, with adaptive logic for small zones.
- Panel Layout Count – Estimates number of panels that can fit on a roof.
- Energy Yield & Financial ROI – Calculates annual kWh, payback years, and carbon offset; uses fallback constants when NASA API is unavailable.
- Performance Heatmap – Visual clusters of low‑yield or neglected zones.
- Executive Briefing – Auto‑generated narrative with Total Zones, Usable Area, Annual Yield, Avg Payback, Carbon Offset.
- Batch Command Center – Upload CSV of 1,000+ coordinates → full city‑wide scan.
- Multi‑Format Export – GeoJSON, Shapefile, KML, CSV, GeoPackage, GeoTIFF, HTML, PDF, DXF (CAD).
- Maintenance Work Order – PDF with fault classification, GPS coordinates, QR codes, and revenue loss.
- Elite UI – Light‑glass themes, animated metrics, pulsing health indicators, panel‑specific diagnostic overlays.
🛠️ INSTALLATION
- Clone & create environment.
- Install dependencies:
pip install -r requirements.txt.
- Download SAM 3 model:
python download_sam3_full.py (requires HF token).
- Verify installation:
python verify_sam3.py.
- Run:
streamlit run web_app.py.
🚀 Performance Expectations
| Area Size | Before (v2) | After (v6 – optimised) |
|---|
| Single roof | 5–8 min | 15–30 sec |
| Neighborhood | 15+ min | 1–2 min |
| City‑wide batch | likely crash | ~8 min (batch mode) |
📄 License
This project uses the SAM 3 model, which is gated. You must accept the license at
https://huggingface.co/facebook/sam3 before downloading.
