Projects

In my MSc reseach, I specifically focused on exploring the potential of directly mapping the degrees of deprivation from Very-High-Resolution (VHR) images. The main work includes 1) developed an integrated methodology of unsupervised learning and deep learning to characterise and quantify deprivation, 2) performed PCA to build data-driven multi-dimensional indices of area deprivation at pixel level and 3) trained a deep CNN-based model via TensorFlow to predict the continuous deprivation indices from VHR images.

The IDEAMAPS Network aims to develop and maintain an Integrated DEprived Area MAPping System that leverages the strengths of our current silo-ed approaches to ‘slum’ area mapping. During my work at the project, I was involved in one of the reseach topics ‘A maximum urban extent definition covering multiple needs’, where I contributed to 1) compared existing urban definitions to arrive at a generalized urban extent conceptualization, 2) co-developed and tested three algorithms to define urban areas in 14 sample cities and 3) worked on a scientific paper together documenting our reseach.

The overall objective of SLUMAP is to propose an open-source framework that allows for the processing of remote sensing images for (i) providing information on the location and extension of slums within a city and (ii) characterizing the physical environment within slums (i.e., in terms of greenness, built-up density, etc.) at limited cost. Our methods will be tested on a set of three sub-Saharan cities containing slums with different characteristics, namely Nairobi (Kenya), Kisumu (Kenya) and Ouagadougou (Burkina Faso).