Airguard VisionEdge

Airguard VisionEdge is a concept for an advanced system that enables environmental researchers and on-site analysts to detect, visualize, and interpret greenhouse gas (GHG) anomalies in near-real-time using a fusion of Edge AI and satellite data.

🌍 Concept Summary

Purpose: To enable environmental researchers and on-site analysts to detect, visualize, and interpret greenhouse gas (GHG) anomalies in near-real-time using Edge AI + satellite fusion.

System Architecture

The system is designed with three core layers:

1. Edge Impulse Node:

   • Runs optimized Machine Learning models (TinyML) for local emission pattern detection.
   • Performs inference on satellite raster tiles and ground sensor data (e.g., $CO_2$, $PM2.5$).

2. Web Dashboard (Android-first):

   • Displays fused insights through interactive maps, charts, and anomaly markers.
   • Syncs with Google Colab notebooks for advanced analytics and deep-dive visualization.

3. LLM Assistant (VisionEdge Copilot):

   • An on-device LLM assistant that explains observed trends.
   • Recommends research insights and provides context for data anomalies.

📱 UX Flow Overview

1. Login & Device Sync Screen

   • Users sign in via Google or their institutional account.
   • Sync connected Edge Impulse devices via Bluetooth/WiFi.
   • The "Add New Station" feature detects and registers a local AI node.

2. Home Dashboard

   • Top Bar: "VisionEdge" title with quick filters (Region | Model | Timeframe).
   • Live Map Panel: Displays raster data tiles with overlay layers for GHG, $NO_2$, and temperature. Edge inferences are highlighted as colored hotspots.
   • Mini Stats Bar: Shows key metrics like Emission Index, Confidence Level, and Anomaly Count. Tapping opens an expanded metrics view.

3. Analysis Panel

   • Organized into tabs: AI Inference | Time Series | Correlations | Ground Data.
   • Features interactive plots generated from Edge outputs.
   • An "Open in Colab" option launches a notebook session with linked data for deeper analysis.

4. Ingestion

   • Upload sensor data files directly to Edge Impulse project for training and analysis.

5. Copilot Assistant

   • A floating chat widget allows users to "Ask VisionEdge Copilot."
   • Users can query the system with natural language, e.g., “Explain today’s emission spike in the Cairo region” to receive an AI-driven explanation.

6. Export & Share

   • Download comprehensive reports as PDF or GeoTIFF files.
   • Push results directly to a shared research group or an institutional drive.

🎨 Design Direction

• Theme: A space black background with green-cyan gradients to represent emission heatmaps.
• UI Style: Sleek and minimal, following Material 3 design principles with a card-based layout.
• Interactions: Smooth map transitions, animated data updates, and collapsible charts for a fluid user experience.
• Data Visualization:
  • 2D raster overlays with opacity controls.
  • Dynamic graphs for comparing local inferences and historical data.

Contributing

Contributions, issues, and feature requests are welcome. For significant contributions, please propose an issue first to discuss what you would like to change.

License

Licensed under the MIT License. See LICENSE for details.

Authors

• [Ahmed Ibrahim Metawee]
• [AIMTY]