Sar Clinometry- and Interferometry-derived Dem Reconstruction

The remote sensing technique has been widely applied in global topographic survey, e.g., for computing a digital elevation model (DEM). The synthetic aperture radar (SAR) is able to perform a synoptic monitoring of the Earth under all-weather and day-and-night conditions. Among the SAR-derived DEM reconstruction techniques, four different methods, i.e., clinometry, grammetry, interferometry and polarimetry, are developed and they are reviewed in [1]. Compared with only one method used in the DEM reconstruction, a combination of several ones can alternatively simplify the procedures or increase the accuracy, e.g., a combination of grammetry and interferometry [2], or a combination of clinometry and polarimetry [3]. In this paper, a novel method is proposed by combining of clinometry and interferometry to improve the accuracy of DEM reconstruction. The technical details of the clinometry method can be referenced in [4], and the interferometry method in [5]. Such a combination is also reasonable if considering the exploitation of both the intensity (clinometry) and phase (interferometry) of the plural SAR signal. From the experimental results with the Envisat data, the interferometry-derived DEM is much more accurate in regions of high coherence than the clinometry-derived result. However, in regions of low coherence, some pronounced errors are remained with the interferometry method due to phase unwrapping problems and other affects. On the other hand, the clinometry method can produce a more robust DEM reconstruction. Therefore a user-defined weighting factor is introduced to optimize the results by using the clinometry-DEM in regions of lower coherence, and the interferometryderived DEM in regions of higher coherence. Finally, the accuracy of the fused DEM can be increased to a great extend.