Identifying and classifying slum areas using remote sensing

Abstract

Thesis taking UN-HABITAT’s definition of slum household as a basis, this study proposes an ontological framework that conceptualizes slums using morphological indicators of the built environment; demonstrates a case of applicability of GSO and explores the transferability of an OOA method for slum classification developed on a single image area to other image areas; proposes a hierarchical method for slum detection by using a combination of texture and spatial metrics; demonstrates a method to quantify the uncertainties of slum interpretations performed by experts; presents the main research findings and conclusions corresponding to each research objective.