Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal HYD0326

TitleAutomatic extraction of water and flood areas from TerraSAR-X data
Investigator Hahmann, Thomas - German Aerospace Center, DFD
Team Member
Mr. Twele, André - German_Aerospace_Center, DFD
Mr. Martinis, Sandro - German_Aerospace_Center, DFD
Prof. Buchroithner, Manfred - TU_Dresden, Institut_fuer_Kartographie
Mr. Roth, Achim - German_Aerospace_Center, DFD
SummaryThe aim of this project is to explore and expand the applicability of high resolution TerraSAR-X data for water and flood mapping. Since flood events are usually accompanied by overcast sky conditions, weather-independent SAR platforms are particularly suitable for obtaining spatially-explicit information about inundated areas in a time- and cost-efficient manner. Past and current C-band SAR platforms (e.g. ERS-2, ENVISAT ASAR) have a proven track-record for large-scale flood mapping. Their potential for deriving flood perimeters in small-scaled scenarios such as urban areas is clearly limited. However, with the launch of the German X-band SAR platform TerraSAR-X which is capable of providing very high spatial resolutions, a detailed assessment of flood extent and damages should be possible. In order to adapt to the demands and capabilities of TerraSAR-X, within the SAR-HQ project object-based and pixel-based processing and analysis techniques are being developed. The project aims towards the integration of SAR data into operational processing and mapping workflows to ensure a fast and reliable access to detailed crisis information. By combining SAR-derived flood maps with additional data sources such as topographic maps and digital elevation data, information products for disaster management, risk assessment and post-hazard reconstruction will be generated. A second key activity within this project is to derive a water body dataset with a global coverage as a contribution to the TanDEM-X project. This upcoming DLR mission shall generate a world-wide, consistent, timely, high-precision Digital Elevation Model with an unprecedented accuracy corresponding to the DTED-3 specifications (12 m posting, 2 m relative height accuracy for flat terrain). According to the TanDEM-X data acquisition concept a minimum of two datasets for every continental area on Earth will be available. Using these datasets both inland waters as well as coastlines shall be identified reliably. Both pixel-based and object-based water detection approaches as well special filter algorithms shall be developed in this project. Different forms of appearance of water in TerraSAR-X imagery shall be analysed. Calm water bodies appear dark in SAR images as the backscatter value is very small due to specular reflection at the water surface. Difficulties in detecting water arise from conditions which represent not this ideal case. The effects of vegetation in the water, sea ice and lake ice, salinisation and aggradation to the SAR signal shall be studied. Furthermore wind induced waves have an important influence on the backscatter values of the radar signal. Ships, which appear very bright in radar images, have to be filtered out of the water body dataset. Ancillary data to detect water bodies in TerraSAR-X images will be the SRTM Water Body Dataset and the C-band as well as X-Band DEM from the SRTM mission.

Back to list of proposals

© DLR 2004-2016