|
1 | | -<div align="center"> |
2 | | -<img width="1200" height="475" alt="GHBanner" src="https://github.com/user-attachments/assets/0aa67016-6eaf-458a-adb2-6e31a0763ed6" /> |
3 | | -</div> |
| 1 | +# Airguard VisionEdge |
4 | 2 |
|
5 | | -# Run and deploy your AI Studio app |
| 3 | +**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. |
6 | 4 |
|
7 | | -This contains everything you need to run your app locally. |
| 5 | +## 🌍 Concept Summary |
8 | 6 |
|
9 | | -View your app in AI Studio: https://ai.studio/apps/drive/1v-6gmO3NbzaoIYMnQoFR10ib-7kjQWrO |
| 7 | +**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. |
10 | 8 |
|
11 | | -## Run Locally |
| 9 | +## System Architecture |
12 | 10 |
|
13 | | -**Prerequisites:** Node.js |
| 11 | +The system is designed with three core layers: |
14 | 12 |
|
| 13 | +1. **Edge Impulse Node:** |
| 14 | + * Runs optimized Machine Learning models (TinyML) for local emission pattern detection. |
| 15 | + * Performs inference on satellite raster tiles and ground sensor data (e.g., $CO_2$, $PM2.5$). |
15 | 16 |
|
16 | | -1. Install dependencies: |
17 | | - `npm install` |
18 | | -2. Set the `GEMINI_API_KEY` in [.env.local](.env.local) to your Gemini API key |
19 | | -3. Run the app: |
20 | | - `npm run dev` |
| 17 | +2. **Web Dashboard (Android-first):** |
| 18 | + * Displays fused insights through interactive maps, charts, and anomaly markers. |
| 19 | + * Syncs with Google Colab notebooks for advanced analytics and deep-dive visualization. |
| 20 | + |
| 21 | +3. **LLM Assistant (VisionEdge Copilot):** |
| 22 | + * An on-device LLM assistant that explains observed trends. |
| 23 | + * Recommends research insights and provides context for data anomalies. |
| 24 | + |
| 25 | +## 📱 UX Flow Overview |
| 26 | + |
| 27 | +1. **Login & Device Sync Screen** |
| 28 | + * Users sign in via Google or their institutional account. |
| 29 | + * Sync connected Edge Impulse devices via Bluetooth/WiFi. |
| 30 | + * The "Add New Station" feature detects and registers a local AI node. |
| 31 | + |
| 32 | +2. **Home Dashboard** |
| 33 | + * **Top Bar:** "VisionEdge" title with quick filters (Region | Model | Timeframe). |
| 34 | + * **Live Map Panel:** Displays raster data tiles with overlay layers for GHG, $NO_2$, and temperature. Edge inferences are highlighted as colored hotspots. |
| 35 | + * **Mini Stats Bar:** Shows key metrics like Emission Index, Confidence Level, and Anomaly Count. Tapping opens an expanded metrics view. |
| 36 | + |
| 37 | +3. **Analysis Panel** |
| 38 | + * Organized into tabs: `AI Inference` | `Time Series` | `Correlations` | `Ground Data`. |
| 39 | + * Features interactive plots generated from Edge outputs. |
| 40 | + * An "Open in Colab" option launches a notebook session with linked data for deeper analysis. |
| 41 | +4. **Ingestion** |
| 42 | + * Upload sensor data files directly to Edge Impulse project for training and analysis. |
| 43 | +5. **Copilot Assistant** |
| 44 | + * A floating chat widget allows users to "Ask VisionEdge Copilot." |
| 45 | + * 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. |
| 46 | + |
| 47 | +6. **Export & Share** |
| 48 | + * Download comprehensive reports as PDF or GeoTIFF files. |
| 49 | + * Push results directly to a shared research group or an institutional drive. |
| 50 | + |
| 51 | +## 🎨 Design Direction |
| 52 | + |
| 53 | +* **Theme:** A space black background with green-cyan gradients to represent emission heatmaps. |
| 54 | +* **UI Style:** Sleek and minimal, following Material 3 design principles with a card-based layout. |
| 55 | +* **Interactions:** Smooth map transitions, animated data updates, and collapsible charts for a fluid user experience. |
| 56 | +* **Data Visualization:** |
| 57 | + * 2D raster overlays with opacity controls. |
| 58 | + * Dynamic graphs for comparing local inferences and historical data. |
| 59 | + |
| 60 | +## Contributing |
| 61 | + |
| 62 | +Contributions, issues, and feature requests are welcome. |
| 63 | +For significant contributions, please propose an issue first to discuss what you would like to change. |
| 64 | + |
| 65 | +## License |
| 66 | + |
| 67 | +Licensed under the MIT License. See `LICENSE` for details. |
| 68 | + |
| 69 | +## Authors |
| 70 | + |
| 71 | +* [Ahmed Ibrahim Metawee] |
| 72 | +* [AIMTY] |
0 commit comments