Humanitarian Satellites : a remote sensing approach to built-up area estimation in refugee camps

Sammanfattning: The determination of the built-up area of a refugee camp is an important task as it can be used for estimating the population. While high resolution satellite imagery can be costly, since 2015 Sentinel-2 data has become available through the European Space Agency. Sentinel-2 offers 10 m resolution images with frequents updates. In this paper three methodologies are compared for the purpose of estimating built-up area using Sentinel-2 imagery: Index Based (using the IBI); Supervised classification; and Digitization. The study areas chosen for this study are Nyarugusu Refugee Camp, in Tanzania, and Kutupalong Refugee Camp, in Bangladesh. Of the three methodologies the Digitization offers the best in the accuracy analysis, followed by the Supervised Classification and, finally, the Index Based Method. Yet, the other two methodologies could still find some specific application such as the delineation of small features. Furthermore, the three methodologies are used in creating a time series analysis to investigate their potentials to track the development of the built-up area of a refugee camp. The built-up area is then correlated to the population of the camp. Once again, the Digitization Method proves itself to be the most accurate of the three methodologies explored.