Calibration of interferometric synthetic aperture radar digital elevation models (DEM) using error surface interpolation methods

Digital elevation models (DEM) are numerical representation of terrain elevation data that has been used in a wide range of spatial analysis applications. The principles for acquisition, storage, management, update, spatial analysis, visualization as well as integration with other systems are reasonably well known. However, as DEM applications are becoming increasingly more widespread, so does concern about the quality of the available elevation data and the propagation of DEM errors through the analysis. It is now well known by analysts and researchers that the results of many DEM-based quantitative operations are significantly affected by the magnitude and also by the spatial distribution of elevation errors in DEM. However, currently available DEM's frequently report only the average magnitude of errors as the root mean square error (RMSE), which does not provide information on systematic bias nor on the spatial patterns of the DEM errors (Heuvelink, 1998). A very important critical point to consider when remote sensing techniques are used for mapping earth resources is the geometric quality and error assessment of data and information products involved throughout the entire process. In the case of 3D data for DEM applications the most suitable and general form to evaluate errors and provide corrections refinement is based in the establishment of a set of mapping functions between feature's positions present in DEM and its corresponding in a cartographic reference frame. So far, many efforts have been made to solve registration and geometric correction problems based in this kind of approach (Audette et al, 2000; Gruen & Akca, 2005). However, relevant research contributions are still required in this field, mainly for development of non-rigid and some degree elastic correction and registration methods, as well as to provide control and certify DEM internal quality (Pottmann et al, 2004). It is an essential issue for topographic mapping applications. The present work deals with DEM generated by synthetic aperture radar (SAR) interferometry. The main objective is developing a method for DEM correction providing statistical estimation of the method's efficiency. The DEM's presently used was generated by SAR P band fully polarimetric system and HH polarization in X band. Data was collected in two-pass mode for P band interferometry and single-pass mode for X band interferometry. SAR radiation in X band do not penetrates into forest so resulting DEM is related to top forest cover, while SAR radiation P band penetrates into forest canopy up to reach ground level, …