Abstract
During humanitarian crises, when population figures are often urgently required but very difficult to obtain, remote sensing is able to provide evidence of both present and past population numbers. This research, conducted on QuickBird time-series imagery of the Zam Zam internally displaced person (IDP) camp in Northern Darfur, investigates automated analysis of the camp's evolution between 2002 and 2008, including delineation of the camp's outlines and inner structure, employment of rule-based extraction for two categories of dwelling units and derivation of population estimates for the time of image capture. Reference figures for dwelling occupancy were obtained from estimates made by aid agencies. Although validation of such ‘on-demand’ census techniques is still continuing, the benefits of a fast, efficient and objective information source are obvious. Spatial, as well thematic, accuracy was, in this instance, assessed against visual interpretation of eight 200 m × 200 m grid cells and accuracy statistics calculated. Total user's and producer's accuracy rates ranged from 71.6% up to 94.9%. While achieving promising results with respect to accuracy, transferability and usability, the remaining limitations of automated population estimation in dynamic crisis situations will provide a stimulus for future research.
Acknowledgements
Work carried out by D. Tiede, D. Hölbling and P. Füreder was funded by the European Commission (EC) within the FP-6 LIMES project (Land and Sea Integrated Monitoring for European Security, Contract No. SIP-CT-2006-031046). S. Lang and P. Zeil have been financed through the FP-7 G-MOSAIC project (GMES Services for Management of Operations, Situation Awareness and Intelligence for Regional Crises, Contract No. 218 822). We would like to thank Olaf Kranz, DLR, coordinator of the Humanitarian Cluster within LIMES, for managing the request of DG RELEX that triggered this research. We would also like to express our thanks to those project partners who have collaborated within this task (EUSC, Joanneum Research, SERTIT, JRC, Definiens AG). Finally, we acknowledge constructive support of Michael Hagenlocher in working out details of the Zam Zam camp during his Masters Thesis.