-orange){kind=icon}
Protecting Chengalpattu District's waterbodies — automatically, 24x7, without manual intervention.
Tamil Nadu is losing its waterbodies at an alarming rate. Encroachments — illegal constructions, land filling, sand mining, and waste dumping — go undetected for months because:
- Traditional systems rely on public complaints, which are often unreported or unnoticed
- Manual inspection is slow, costly, and geographically limited
- By the time authorities act, the damage is irreversible
The consequences are devastating: flooding during heavy rains, reduced groundwater recharge, and permanent loss of water sources for future generations.
ArenIQ is a dual-layer encroachment detection and reporting system:
| Layer | What it does |
|---|---|
| 🛰️ Satellite Monitoring Website | Automatically detects changes in waterbodies using Sentinel-2 imagery and sends alerts to authorities |
| 📱 Citizen Reporting App | Allows verified citizens to photograph and report encroachments with GPS tagging |
Together, they form a dual authentication + complaint escalation system that bridges the gap between detection and government action.
┌─────────────────────────────────────────────────────┐
│ ArenIQ System │
│ │
│ ┌──────────────────────┐ ┌──────────────────────┐ │
│ │ 🛰️ Satellite Module │ │ 📱 Citizen App │ │
│ │ │ │ │ │
│ │ Sentinel-2 │ │ Flutter (Android/ │ │
│ │ (Copernicus/ESA) │ │ iOS) │ │
│ │ ↓ │ │ ↓ │ │
│ │ NDWI Calculation │ │ OTP Verification │ │
│ │ ↓ │ │ ↓ │ │
│ │ Image Differencing │ │ GPS-Tagged Report │ │
│ │ ↓ │ │ ↓ │ │
│ │ Change Detection │ │ Supabase Storage │ │
│ │ ↓ │ └──────────────────────┘ │
│ │ Random Forest │ │ │
│ │ Classifier │ │ │
│ └──────────────────────┘ │ │
│ │ │ │
│ └────────────┬────────────┘ │
│ ↓ │
│ ┌───────────────────────┐ │
│ │ Node.js Backend │ │
│ │ Alert Engine │ │
│ │ Escalation Logic │ │
│ └───────────────────────┘ │
│ ↓ │
│ ┌───────────────────────┐ │
│ │ Ntfy.sh Push Alerts │ │
│ │ → Local Authority │ │
│ │ → District Level │ │
│ │ → State Level │ │
│ └───────────────────────┘ │
└─────────────────────────────────────────────────────┘
Step 1 — Image Acquisition
Satellite images are fetched from Sentinel-2 via the Copernicus Open Access Hub using the open-source sentinelsat Python library. Images are pulled on a daily/weekly schedule for all registered waterbodies in Chengalpattu District.
Step 2 — NDWI Calculation The Normalized Difference Water Index is computed for both current and previous images:
NDWI = (Green − NIR) / (Green + NIR)
Higher NDWI values indicate water presence. A drop in NDWI signals potential encroachment.
Step 3 — Image Differencing The old NDWI map is subtracted from the current one to isolate only the changed regions, filtering out seasonal variation noise.
Step 4 — Change Area Extraction Pixels showing significant change (beyond a calibrated threshold) are grouped into contiguous zones — these are candidate encroachment sites.
Step 5 — Encroachment Classification A Random Forest Classifier (scikit-learn) trained on labeled satellite data categorizes each detected change:
- 🏗️ Construction / Building
- 🪨 Sand Mining
- 🗑️ Waste Dumping
- 🌍 Land Filling
Step 6 — Alert Dispatch A complaint is auto-generated and sent via Ntfy.sh (open-source, self-hostable push notifications) to the responsible local authority. If no action is taken within the deadline, the alert automatically escalates to higher officials.
Step 1 — Verified Sign-Up Users register with their mobile number and verify via OTP (Supabase Auth). This prevents fake or anonymous reports.
Step 2 — Report Encroachment The user taps "Report Encroachment", captures or uploads a photo, and optionally adds a short caption describing what they observed.
Step 3 — GPS Tagging Location is automatically captured via the device GPS or manually pinned on an OpenStreetMap + flutter_map interface. The location determines which authority receives the complaint.
Step 4 — Encroachment Type Selection Users select the type of encroachment from a predefined list (Construction, Dumping, Sand Mining, etc.) for faster processing.
- ❌ No public comments or sharing of reports
- ❌ No screenshots allowed within the app
- ✅ Reports go directly and only to the mapped authority
Unlike research tools like SandWatch or academic satellite platforms where data rarely reaches authorities:
| Feature | ArenIQ | Existing Tools |
|---|---|---|
| 24x7 automated monitoring | ✅ | ❌ |
| Direct authority alert | ✅ | ❌ |
| Escalation to higher officials | ✅ | ❌ |
| Citizen + satellite dual layer | ✅ | ❌ |
| Real-time encroachment count (public) | ✅ | ❌ |
| Encroachments rescued counter | ✅ | ❌ |
| Fully FOSS stack | ✅ | ❌ |
✅ All tools are free and open-source (FOSS). No proprietary APIs required for core functionality.
| Component | Technology | License |
|---|---|---|
| Language | Python | PSF |
| Satellite Data | Sentinel-2 via sentinelsat + Copernicus Open Access Hub |
Apache 2.0 / ESA Open |
| Image Processing | OpenCV | Apache 2.0 |
| NDWI + Change Detection | NumPy, Rasterio | BSD |
| Classification | scikit-learn (Random Forest) | BSD |
| Frontend | React.js | MIT |
| Backend | Node.js | MIT |
| Push Alerts | Ntfy.sh (self-hostable) | Apache 2.0 |
| Component | Technology | License |
|---|---|---|
| Framework | Flutter (Android & iOS) | BSD |
| Maps | OpenStreetMap + flutter_map | ODbL / BSD |
| Authentication (OTP) | Supabase Auth | Apache 2.0 |
| Database | Supabase (PostgreSQL) | Apache 2.0 |
| Image Storage | Supabase Storage | Apache 2.0 |
| Push Notifications | Ntfy.sh | Apache 2.0 |
ArenIQ/
├── Web/ # React.js frontend (satellite monitoring dashboard)
├── App/ # Flutter mobile application
├── backend/ # Node.js backend + alert escalation engine
├── ndwi_detection.py # Core NDWI + change detection + RF classifier
├── LICENSE # MIT License
└── README.md
- Python 3.8+
- Node.js 18+
- Flutter SDK
- Copernicus Open Access Hub account (free — https://scihub.copernicus.eu)
- Supabase project (free tier — https://supabase.com)
cd backend
npm install
cp .env.example .env # Fill in your Supabase credentials
npm startpip install sentinelsat opencv-python scikit-learn numpy rasterio
python ndwi_detection.pycd Web
npm install
npm run devcd App
flutter pub get
flutter runArenIQ is currently focused on Chengalpattu District as a pilot region, chosen because:
- Rapid urban expansion makes encroachment highly active and measurable
- Dense network of registered waterbodies provides rich test data
- Proximity to Chennai makes authority coordination feasible for real-world validation
The system is intentionally scoped to one district to ensure depth over breadth — accurate detection, reliable alerts, and real authority response.
- 🌊 Monitor all registered waterbodies in Chengalpattu District simultaneously
- 🏘️ Reduce urban flooding by catching encroachments before they become permanent
- ⚡ Cut detection-to-action time from months to hours
- 👁️ Provide transparent public accountability through live encroachment + rescue counters
- Small encroachment detection — Sentinel-2's 10m resolution may miss very small-scale changes
- Seasonal waterbody disappearance — Small waterbodies may dry up in summer; handled by per-season baseline calibration
- Citizen adoption — Encouraging consistent public participation requires awareness campaigns
Portions of this codebase and documentation were developed with assistance from Claude (Anthropic). All architecture decisions, domain-specific logic, and implementation were designed and validated by the team.
This project is licensed under the MIT License — see the LICENSE file for details.
Built with ❤️ for Tamil Nadu's waterbodies at Foss Hack 2026.
"Water is not just a resource — it is life. It is our duty to protect it."
| Member | GitHub |
|---|---|
| Mithra J | @Mithra-J |
| Mohammed Faiz Y | @FAIZ1409 |
| Mohamed Marzuq Tharif | @mohamedmarzuqtharif |