michaelakandwanahorugomoka/usip-uganda

# 🛰️ Uganda Satellite Intelligence Platform (USIP) > **An AI-driven Earth observation and environmental monitoring platform utilizing multispectral satellite imagery and machine learning to assess agricultural crop health across Uganda.** --- ## 📌 Overview The **Uganda Satellite Intelligence Platform (USIP)** integrates multi-sensor satellite imagery and meteorological datasets to detect, monitor, and predict localized crop stress in Uganda. By coupling real-time satellite data from **Google Earth Engine** with a trained **Random Forest classification model**, USIP delivers actionable agricultural metrics through a high-performance **FastAPI** backend and an interactive **Streamlit GIS dashboard**. --- ## 🏗️ System Architecture ```text [ Sentinel-2 L2A ] [ CHIRPS Rainfall ] [ ERA5-Land Temp ] │ │ │ └─────────────────────────┼─────────────────────────┘ ▼ [ Google Earth Engine Pipeline ] │ ▼ [ Feature Fusion Engine ] (NDVI, Precipitation, Temp) │ ▼ [ Random Forest ML Engine ] (model.joblib) │ ▼ [ FastAPI REST Backend ] (127.0.0.1) │ ▼ [ Streamlit Spatial Dashboard ] (localhost) ``` --- ## ✨ Key Features * **Multi-Sensor Fusion:** Automatically ingests high-resolution **Sentinel-2** (NDVI vegetation index), **CHIRPS** (daily rainfall aggregates), and **ERA5-Land** (surface air temperature). * **Predictive ML Classification:** Evaluates environmental indicators to classify crop status into three health tiers: * 🟢 **Healthy** (`NDVI > 0.60`) * 🟡 **Moderate Stress** (`0.40 <= NDVI <= 0.60`) * 🔴 **High Stress** (`NDVI < 0.40`) * **REST API Service:** Built on **FastAPI** for real-time spatial inference and integration with external GIS clients. * **Interactive GIS Dashboard:** Interactive monitoring across Ugandan districts (Kamwenge, Kasese, Gulu, Mbarara, Mbale) powered by **Streamlit** and **Folium**. --- ## 🔌 API Documentation ### `GET /api/v1/assess-crop-health` …