Object-oriented analysis of very high resolution orthophotos for estimating the population of slum areas , case of dares-salaam , tanzania

Obtaining up-to-date spatial information about slum settlements is of great importance for decision making related to slum and housing policy. The availability of Very High Resolution (VHR) satellite imagery in meter or sub-meter level is helpful and promising in the information extraction of slum areas based on object-oriented techniques. This paper aims to determine the feasibility of using VHR orthophotos (0.6 meter pixels) to create an accurate inventory of buildings in order to estimate the slum population. eCogniton software is used for the image segmentation and classification of the objects of interest (building roofs) for three different slum areas in Dar-Es-Salaam called “Charambe”, “Manzese” and “Tandale”. Roof extraction accuracy is computed on the study wards. In total, out of 1650 reference buildings 1504 buildings are extracted (91% accuracy). The number of inhabitants living in a dwelling unit is obtained from household surveys and used to calculate the Roof Area per Person (RApP) rate. The estimated population by the applied model represents 82.2%, 72.5% and 68.3% for Charambe, Manzese and Tandale wards, respectively. * Corresponding author. Email: mehdiaminipour@gmail.com