abel-tesfaye-geo/GEE-Supervised-Classification

Land use land cover classification using supervised machine learning in Google Earth Engine for the city Bahir Dar, Ethiopia # GEE-Supervised-Classificatio ## Supervised Machine Learning | Google Earth Engine ## Overview This project performs supervised image classification to map land use and land cover (LULC) in [your study area], Ethiopia, using Google Earth Engine (GEE). The classification uses Sentinel-2 / Landsat imagery and a Random Forest classifier. ## Study Area - Location: Bahir Dar, Ethiopia - Area: 600 km² - Imagery: Sentinel-2 / Landsat 8 (Year: 2024) ## Classification Classes - Urban / Built-up - Agricultural Land - Forest / Vegetation - Waterbody - Bare Land ## Methodology 1. Image acquisition and cloud masking 2. Training sample collection 3. Feature extraction (spectral bands + indices) 4. Random Forest classification 5. Accuracy assessment ## Results ## Accuracy Assessment | Class | Producer's Accuracy | User's Accuracy | |-------|-------------------|----------------| | Urban |97 | % | | Agriculture |89 | % | | Forest |90 | % | | Water 92 | % | | Bare Land | 93 | % | **Overall Accuracy:** 90% **Kappa Coefficient:** 0.89 ## Code The full GEE script is available in the `code/` folder. ## Tools Used - Google Earth Engine (JavaScript API) - Sentinel-2 / Landsat 8 imagery - QGIS (for visualization) ## Author Abel Tesfaye Assistant Lecturer, Institute of Land Administration Bahir Dar University, Ethiopia