Comparison of ALOS-PALSAR and TerraSAR-X Data in Terms of Detecting Settlements – First Results

Thomas Esch(1), Achim Roth(1), Michael Thiel(2), Michael Schmidt(1) and Stefan Dech(1)

(1) German Aerospace Center (DLR), Oberpfaffenhofen, D-82234 Wessling, Germany
(2) University of Würzburg, Am Hubland, D-97074 Würzburg, Germany


The rapid increase of urban population in the last decades has been attended by a constantly accelerating urban sprawl. In order to cope with the dynamics of urban agglomerations resource managers, environmentalists and politicians require up-to-date information on the characteristics and changes of urban environments. Spaceborne earth observation (EO) has become a promising tool to provide such geo-information. In previous studies we could demonstrate the applicability of TerraSAR-X data for an accurate detection of built-up areas. The according methodology includes a specific pre-processing and an object-oriented classification of the SAR images. The pre-processing includes an enhanced speckle analysis and suppression providing an intensity image with minimized speckle and a texture layer accentuating urban areas. Based on the filtered intensity data and the textural information the classification process starts with the generation of segments. The resulting image objects are used for the extraction of temporally and spatially robust hierarchical and contextual features that afford an accurate delineation of urban footprints. This paper presents first results of a transfer of this methodology to ALOS-PALSAR images. We could show that the basic procedure is applicable to ALOS L-band images. However, so far the accuracy of an according analysis based on ALOS data is lower compared to the use of X-band images. Hence, further modifications are required in order to improve the presented methodology for L-band data.


Symposium presentation


  Higher level                 Last modified: 07.05.06