A Spatiotemporal Ontology of Informal Settlements Using a Combination of OBIA-RF With Worldview-3 and Landsat Data

An understanding of the spatial distribution of informal settlements within a city is important for urban management decision-making and service infrastructure provision, provides useful information for planners and policymakers, and has a role in minimizing future urban environmental issues. The objective of this article is to evaluate the performance of an ontology of informal settlements mapping for Riyadh city. Satellite data include a combination of medium-resolution Landsat thematic mapper, enhanced thematic mapper plus, and operational land imager, and VHR Worldview-3 imagery. Object-based image analysis (OBIA) technique was employed to identify 30 useful indicators at defined object, settlement, environment, and temporal levels. Time-series analysis (TSA) was undertaken, and a multidimensional model was developed to define the trend of changes through 30 years. The classification process incorporated OBIA, random forest (RF), and LandTrendr techniques. The classification output included delineation of formal and informal settlement boundaries and road networks, as well as vegetated and vacant areas. The final OBIA-RF and TSA classification demonstrated an overall accuracy (OA) of 89% with the corresponding kappa value of 87%. The OBIA-RF classification developed without TSA techniques returned an OA of 87% and kappa value of 84%. The article indicated that using OBIA and RF methods, in combination with LandTrendr, can be a useful tool for planners and decision-makers to identify changes in the land cover of informal settlements within Riyadh city and beyond.