With rapid urbanization leading to the proliferation of deprived urban areas (often referred to as “slums”) in sub-Saharan
Africa, there is a growing number of city dwellers living in inadequate housing conditions and being exposed to multiple hazards.
However, there are huge data gaps in geospatial data on these areas for informing stakeholders and decision-makers on different
levels, form local to international. The research project SLUMAP is an effort to develop an open-source framework with suitable methods
and tools for mapping and characterising deprived areas in sub-Saharan Africa (SSA), considering costs, benefits, and user
information requirements. Our methods are being tested on three cities, namely Nairobi (Kenya), Kisumu (Kenya) and Ouagadougou (Burkina
Faso).
At the city scale, we assess and compare the merits of different free or low-cost image datasets (Sentinel-1/2, SPOT...)
for mapping the location and extent of deprived areas. We develop automated state-of-the-art workflows based on free open-source
software (FOSS), in view of favouring the scalability and transferability of the approaches. We exploit and combine a vast array of
spectral, spatial and morphological features and identify from them the best predictors, for limiting data and processing costs in
applications, while maintaining a high level of accuracy. To avoid stigmatizing neighbourhoods and exposing vulnerable
populations, we present the results as gridded maps of morphological deprivation probability. Besides, focusing on each deprived area, we
employ very-high resolution super-spectral WorldView-3 imagery (with several combinations of XS and SWIR bands) in an adapted
GEOBIA/Machine Learning processing workflow to produce LULC classifications that include ad hoc classes (such as ‘Garbage piles’),
and subsequently derive metrics designed to characterize deprived areas. Here also, we assess and draw conclusions on trade-offs
for meeting user requirements while reducing data and processing costs.
Web site: https://slumap.ulb.be/