Print

Joint Topic: Remote Sensing Methodologies

In the Civil Crisis Information and Geo-Risks department, methods based on remote sensing are developed and implemented for all parts of the disaster cycle, from prevention and risk minimization to early detection and immediate disaster response to reconstruction. On the one hand these methods aim at timely derivation of crisis information after a catastrophe (for example, satellite-supported detection of flooded or burning areas), and on the other hand analytical tools are developed which facilitate early warning or risk assessment before a catastrophe develops.

Close contact among those who develop methodologies in the department is essential in order to assure networking and an exchange of information as insights are applied. Working in a common development environment and defining uniform software development standards especially supports the continuity and stabilization of methodology development.

The team focuses on developmental work which includes the following:

• Contextual classification methodologies for SAR satellite data created by combining causal and noncausal Markov imaging models on hierarchically irregular graphs
• Methods for early detection over large areas of flooding on a continental scale with the help of satellites
• Methods of interactive pixel aggregation in high resolution optical satellite images using graph-based modeling
• Modeling empirical and heuristic image characteristics.

These developments are implemented as semi- and fully-automated image processing tools and processing chains which can be used to detect crisis-relevant phenomena in optical and radar satellite data, and, as needed, to follow the course of their temporal evolution. The work of this team supports such activities as the routine operations of the Center for Satellite-based Crisis Information (ZKI).