TerraSAR-X Radargrammetry Products From Airbus Defence and Space

The University of Miami group received first radargrammetry results from industry partner Airbus Defence and Space in December. These land topography products are based on a set of SAR images of the area on the Florida panhandle where Hurricane Michael made landfall in 2018. The test images were acquired between August 26 and September 1, 2021, as reported earlier.

Radargrammetry applies the principles of photogrammetry / stereo imaging to SAR images, which means that two or more SAR images acquired from different angles are used to determine the vertical locations of objects from their relative horizontal shifts between the images. This is different from SAR interferometry, which uses two images acquired from almost the same angle and determines heights from phase differences between complex pixel amplitudes. In that case, the two view angles must be so close to each other that there are no visible shifts. While SAR interferometry is usually more accurate than radargrammetry, the acquisition of a suitable pair of SAR images with a single satellite can take many days (a full 11-day repeat cycle in the case of TerraSAR-X) or be not feasible at all if the satellite is not on a repeat orbit. Image pairs for radargrammetry can be collected within a much shorter period of time. Airbus has experience with radargrammetry because the technique has been used to fill gaps in the WorldDEM (TanDEM-X) data set in areas where interferometry was not possible. The test images of the Hurricane Michael landfall area were acquired in order to test the methodology and evaluate the quality of the results.

A first look at the provided radargrammetry products for the relatively flat Hurricane Michal test area shows a combination of expected topography variations and pronounced artifacts, which look like noisy long waves in most areas. The animated GIF on this page shows a visual comparison of the topography around Tyndall AFB as seen in the Airbus data set and in a lidar-based reference topography from the NOAA National Centers for Environmental Information (NCEI). For better visibility of the differences, the terrain height variations have been increased by a large factor relative to the horizontal scales. The UMiami team will try in the coming weeks to remove the processing artifacts from the Airbus data sets by post processing. In parallel, the partners at Airbus will look into possibilities to improve their algorithm for this specific application and reduce the magnitude of processing artifacts from the beginning.