Data4Human (MF-PBA)

The aim of the “Demand-driven Data Services for Humanitarian Aid” (Data4Human) project was to make geo-information from remote sensing data more usable for humanitarian relief missions and to tailor methods to the needs of humanitarian organizations. The focus of MF-PBA was on the analysis, development and provision of methods and processing environments for detecting changes in infrastructure and buildings. The outcome are three deep learning based approaches for extracting roads and buildings and assessing their damage. The methods were evaluated and tested at four test sites in Germany, Indonesia, Mozambique, and Lebanon and were able to achieve reliable results. The development took place in close cooperation with a team from the World Food Program (WFP), which will be continued in the coming years.