Correction of Positional Errors and Geometric Distortions in Topographic Maps and DEMs Using a Rigorous SAR Simulation Technique

In the history of surveying and mapping, a large volumes of topographic maps and digital elevation models have been created at various scales throughout the world. However, positional errors and geometric distortions may exist in the topographic contour maps and their derived DEMs due to inaccurate ground control and poor navigation techniques in the early years. In this paper, we present a new technique to detect and correct positional errors and geometric distortions in topographic data based on rigorous Synthetic Aperture Radar (SAR) image simulation and mathematical modeling of SAR imaging geometry. Our method has been successfully applied to two USGS topographical data sets in Antarctica. Using Radarsat SAR imagery, positional errors of these two data sets have been reduced from 5 km to 200 m and from 200 m to 50 m, respectively. Introduction Traditionally, terrain is depicted by topographic maps with contour lines. Over the last hundred years, numerous topographic maps at various scales have been produced by various mapping agencies, such as the USGS (USGS, 1999). With rapid advances in computer technology, the raster-based Digital Elevation Model (DEM) has replaced topographic maps for collecting, storing, and conveying surface topography information. Recent decades have witnessed a rapid growth in DEM production. Due to their low-cost and wide availability, the contour data digitized from topographic maps were often used as the primary input data source, for example, USGS DEM products (USGS, 1987 and 2000). Consequently, positional errors and geometric distortions that exist in the early topographic maps were propagated into the DEMs. The presence of positional errors and geometric distortions often causes geographic misplacement of topographic features, inhibits the alignment and co-registration of topographic data with other geo-spatial data layers, and hence, plagues subsequent spatial data integration and modeling. The development of the global positioning system (GPS), inertial navigations system (INS), and new topographic data Correction of Positional Errors and Geometric Distortions in Topographic Maps and DEMs Using a Rigorous SAR Simulation Technique Hongxing Liu, Zhiyuan Zhao, and Kenneth C. Jezek acquisition techniques, such as digital photogrammetry based on automated image matching (Miller and Helava, 1992; Heipke, 1992; Mikhail, et al., 2001), Interferometric SAR (INSAR) (Goldstein, et al., 1993; Zebker, et al., 1994) and Light Detection and Ranging (LIDAR) systems (Hill, et al., 2000; Flood, 2001) technologies, have greatly improved the precision of planimetric positioning, as well as, the vertical accuracy of newly acquired elevation measurements. However, to fully utilize the existing topographic data, it is desirable to develop a technique to handle positional errors and geometric distortions. This paper presents a new method for detecting and correcting positional errors and geometric distortions in the existing topographic contour maps and their derived DEMs based on a Synthetic Aperture Radar (SAR) image simulation technique. With Radarsat-1 SAR image data captured in 1997, we analyzed two topographic data sets in the Ellsworth Mountains and Dry Valley regions, Antarctica, which were originally produced by the USGS in the 1960s and 1970s using aerial photogrammetric techniques. Our analysis reveals a planimetric positional error as large as 5 km for topographic features. It was found that the positional shifts are not systematic, and consequently, the shape of the terrain was geometrically distorted. Using rigorous SAR image simulation and mathematical modeling of SAR imaging geometry, we successfully corrected the positional errors and geometric distortions in these topographical data sets. In the following sections, we first describe the algorithms and processing steps involved in our method, then, demonstrate the effectiveness of our methods through two application examples, and finally, present some technical remarks and conclusions.