GAME.EO

Recent years have brought tremendous advancements in the area of automated information extraction from Earth Observation (EO) imagery, but problems still remain since even state-of-the-art algorithms based on imagery alone do not provide a satisfactory solution. In these situations, it is possible to exploit the crowdsourcing of human intelligence, which is a recent promising area for EO. This is of particular interest with respect to providing information on devleoping countries to International Finance Institutions such as the World Bank.In this project an integrated (hybrid) crowdsourced and EO data-based information extraction framework is being developed. Mobile-based tools for supporting crowdsourcing campaigns and gaming approaches will be developed, and then used to mobilize and train volunteers to provide data via dedicated EO-based workflows to extract the required information in a more timely and accurate manner, with lower costs than would be incurred using professional datacollection services.
The approach will be demonstrated using specific service cases for EO-based monitoring of Informal Settlements/Slum Areas (SDG11), with the aim to enhance current machine-learning algorithms for the identification, delineation and further characterization of these areas. The developed framework and tools will be tested in cooperation with World Bank users and stakeholders (GWASP/GSURRP) in an ongoing internal project1 for Dhaka, Bangladesh, to demonstrate the potential and the added value of the synergies of crowdsourcing- and EO-based information to support the World Bank’s research and operational activities.