Google Earth Engine for Informal Settlement Mapping: A Random Forest Classification Using Spectral and Textural Information

Accurate and reliable informal settlement maps are fundamental decision-making tools for planning, expediting informed management of cities. However, extraction spatial information settlements has remained a mammoth task due to the heterogeneity urban landscape components, requiring complex analytical processes. To date, use Google Earth Engine platform (GEE), with cloud computing prowess, provides unique opportunities map precision enhanced accuracy. This paper leverages cloud-based techniques within GEE integrate spectral textural features accurate location extent in Durban, South Africa. The aims investigate potential advantages GEE’s innovative image processing precisely depict morphologically varied settlements. Seven data input models derived from Sentinel 2A bands, band-derived texture metrics, indices were investigated through random forest supervised protocol. main objective was explore value different combinations accurately mapping results revealed that classification based on bands + yielded highest identification accuracy (94% F-score). addition decreased Our confirm is achieved ‘textural features’ model, which lowest root-mean-square log error (0.51) mean absolute percent (0.36). approach highlights capability integrative capabilities extracting morphological variations rugged heterogeneous landscapes,