Real-time air quality estimation using your phone camera and atmospheric optics — works offline, no accounts needed.
Requires camera permission. Torch/flash will only activate with your explicit tap.
ℹ️ About accuracy & limitations
AQI Vision uses atmospheric optics algorithms (Dark Channel Prior, Laplacian sharpness, colour statistics, temporal motion filter) to estimate PM2.5 and PM10 from camera images. The app auto-detects indoor vs. outdoor scenes.
Expected accuracy — Outdoor: Category-level (Good / Moderate /
Unhealthy) ~70–80% reliable. Exact µg/m³ values ±30% uncertainty.
Expected accuracy — Indoor: Category-level ~60–70% reliable.
Works for cooking smoke, dust, incense, poor ventilation.
🌤️ Outdoor — best results when:
• Pointing at open sky or distant landscape
• Adequate daylight, phone held steady 5+ sec
• No glass/window between camera and atmosphere
🏠 Indoor — best results when:
• Lights on, room well-lit (no direct glare)
• Visible room depth ≥ 3 m in frame (far wall / furniture)
• Point away from bare walls and light sources
Not reliable for:
• Very dark, overexposed, or rapidly moving scenes
• Health-critical or regulatory decisions
• High-humidity conditions (fog, mist inflate DCP estimates)
For certified readings use official monitoring stations or calibrated optical particle counters.
Was this AQI data accurate for your area?