Combining Object-Based Machine Learning with Long-Term Time-Series Analysis for Informal Settlement Identification

Informal settlement mapping is essential for planning, as well resource and utility management. Developing efficient ways of determining the properties informal settlements (when, where, who) critical upgrading services planning. Remote sensing data are increasingly used to understand built environments. In this study, we combine two sources data, very-high-resolution imagery time-series Landsat identify describe settlements. The indicators characterising were grouped into four different spatial temporal levels: environment, settlement, object time. These then in an object-based machine learning (ML) workflow proposed method had a 95% overall accuracy at Among levels examined, contribution level was most significant ML model, followed by object-level indicators. Whilst did not contribute greatly classification settlements, it provided way understanding when formed. adaptation would allow combination wide-ranging diverse group comprehensive framework.