What went into this project?
This project focuses on analyzing land cover changes in Khartoum, Sudan between 2013 and 2025 using supervised image classification techniques in ArcGIS Pro.
The workflow started with downloading multi-temporal Landsat satellite imagery from the USGS Earth Explorer for the selected years. The images were preprocessed by stacking spectral bands to create multi-band composites, followed by mosaicking multiple scenes to ensure full coverage of the study area. The mosaicked rasters were then clipped to the exact boundary of the study area.
Different band combinations (false color composites) were applied to enhance feature differentiation and improve class separability. Based on these composites, training samples were created for key land cover classes such as:
Agriculture lands
Built-up areas
Desert/bare land
Green areas
Water bodies
Using these training samples, supervised classification was performed to generate land cover maps for each year.
The entire workflow was executed twice (for 2013 and 2025) using the same methodology to ensure consistency. Finally, the classified outputs were compared to detect and quantify land cover changes over time. The results were visualized through maps and statistical charts to highlight urban expansion, vegetation changes, and shifts in land use patterns.
