Search

A well working transportation infrastructure is crucial for modern societies and the mobility of people. The general public relies on a sound infrastructure to fulfill their increasing and individual mobility needs. This means a wide range of transport necessities – e.g. daily way to work and daily (consumer, service) needs. In the event of a crisis, e.g. caused by a natural hazard, the transport infrastructure becomes vital for different reasons. The people have to be evacuated or rescued from the affected area. Furthermore, relief aid and supplies, as well as rescue forces do require unimpaired access to the crisis area. Natural disasters, like an earthquake or a flooding, are disturbing the transport infrastructure negatively. Cascading consequences are, for example, impassable roads, disturbed railway connections and a reduced transport capacity of persons and goods. To develop evacuation strategies, to plan logistics operations of emergency forces in the affected area and to safely route these forces through the affected area, among others, routing tools are required that take into account the current state of the infrastructure [1]. The Institute of Transportation Systems of the German Aerospace Center has developed KeepOperational, a web-based, integrated decision support tool for traffic management and crisis logistics. As a key feature, this system can directly inject impairments of the transportation infrastructure – derived from various sources – into the routing network so that these disturbances will be considered by the routing algorithms.

In this paper we show the pre-processing and potential for environmental applications of very high resolution (VHR) satellite stereo imagery like these from WorldView-2 or Pl´eiades with ground sampling distances (GSD) of half a metre to a metre. To process such data first a dense digital surface model (DSM) has to be generated. Afterwards from this a digital terrain model (DTM) representing the ground and a so called normalized digital elevation model (nDEM) representing off-ground objects are derived. Combining these elevation based data with a spectral classification allows detection and extraction of objects from the satellite scenes. Beside the object extraction also the DSM and DTM can directly be used for simulation and monitoring of environmental issues. Examples are the simulation of floodings, building-volume and people estimation, simulation of noise from roads, wave-propagation for cellphones, wind and light for estimating renewable energy sources, 3D change detection, earthquake preparedness and crisis relief, urban development and sprawl of informal settlements and much more. Also outside of urban areas volume information brings literally a new dimension to earth oberservation tasks like the volume estimations of forests and illegal logging, volume of (illegal) open pit mining activities, estimation of flooding or tsunami risks, dike planning, etc. In this paper we present the preprocessing from the original level-1 satellite data to digital surface models (DSMs), corresponding VHR ortho images and derived digital terrain models (DTMs). From these components we present how a monitoring and decision fusion based 3D change detection can be realized by using different acquisitions. The results are analyzed and assessed to derive quality parameters for the presented method. Finally the usability of 3D information fusion from VHR satellite imagery is discussed and evaluated.