Time-series satellite imagery analysis to track deforestation and generate ranger alerts across Cross River, Congo Basin and Niger Delta.
# Deforestation & Land Use Change Monitoring
Time-series satellite imagery analysis to track illegal logging, land encroachment, and deforestation across Cross River State, Congo Basin, and Niger Delta, generating automated alerts for forest rangers and policy makers.
---
## Problem Statement
Nigeria loses ~350,000 hectares of forest annually, one of the world's worst deforestation rates. The Niger Delta mangroves and Cross River rainforests are critically threatened by illegal logging and agricultural encroachment. This system provides near-real-time detection and alert generation.
---
## Features
| Feature | Description |
|---------|-------------|
| Multi-Year NDVI Analysis | Time-series 2015–2024 from Landsat-8 / Sentinel-2 |
| Land-Use Classification | Random Forest, Forest / Vegetation / Agriculture / Deforested |
| Automated Alert Generation | Critical / High / Moderate severity alerts for rangers |
| Forest Coverage Trends | Annual loss rates by region and forest type |
| Change Detection Map | Interactive map with deforestation alert markers |
---
## Tech Stack
| Layer | Technology |
|-------|-----------|
| Remote Sensing | Landsat-8, Sentinel-2 NDVI/EVI |
| Machine Learning | scikit-learn (Random Forest) |
| Geospatial | GeoPandas, Folium, Rasterio |
| Visualisation | Matplotlib, Seaborn, Plotly |
| Data | pandas, NumPy |
---
## Project Structure
```
deforestation-monitoring/
├── src/
│ ├── data_loader.py # Time-series satellite data ingestion and alert detection
│ ├── model.py # Land-use classification, forest loss computation
│ └── visualize.py # Change map, trend charts, land-use pie charts
├── data/raw/ # Landsat/Sentinel rasters, shapefiles
├── models/ # Saved Random Forest classifier
├── config.yaml # NDVI thresholds, alert levels, region config
├── main.py # Pipeline entry point
└── requirements.txt
```
---
## Quick Start
```bash
git clone
githu …