Accuracy analysis for DSM and orthoimages derived from SPOT HRS stereo data using direct georeferencing

Abstract During the HRS (High Resolution Stereo) Scientific Assessment Program the French space agency CNES delivered data sets from the HRS camera system with high precision ancillary data. Two test data sets from this program were evaluated: one is located in Germany, the other in Spain. The first goal was to derive orthoimages and digital surface models (DSM) from the along track stereo data by applying the rigorous model with direct georeferencing and without ground control points (GCPs). For the derivation of DSM, the stereo processing software, developed at DLR for the MOMS-2P three line stereo camera was used. As a first step, the interior and exterior orientation of the camera, delivered as ancillary data from positioning and attitude systems were extracted. A dense image matching, using nearly all pixels as kernel centers provided the parallaxes. The quality of the stereo tie points was controlled by forward and backward matching of the two stereo partners using the local least squares matching method. Forward intersection lead to points in object space which are subsequently interpolated to a DSM in a regular grid. DEM filtering methods were also applied and evaluations carried out differentiating between accuracies in forest and other areas. Additionally, orthoimages were generated from the images of the two stereo looking directions. The orthoimage and DSM accuracy was determined by using GCPs and available reference DEMs of superior accuracy (DEM derived from laser data and/or classical airborne photogrammetry). As expected the results obtained without using GCPs showed a bias in the order of 5–20 m to the reference data for all three coordinates. By image matching it could be shown that the two independently derived orthoimages exhibit a very constant shift behavior. In a second step few GCPs (3–4) were used to calculate boresight alignment angles, introduced into the direct georeferencing process of each image independently. This method improved the absolute accuracy of the resulting orthoimages and DSM significantly.

[1]  Peter Reinartz,et al.  From Airborne Digital Raw Data to Image Maps , 2003 .

[2]  E. Breton,et al.  SPOT5 geometric image quality , 2003, IGARSS 2003. 2003 IEEE International Geoscience and Remote Sensing Symposium. Proceedings (IEEE Cat. No.03CH37477).

[3]  A. Roth,et al.  The shuttle radar topography mission—a new class of digital elevation models acquired by spaceborne radar , 2003 .

[4]  Achim Roth,et al.  Evaluation of Interferometric Digital Elevation Models Derived from ERS Tandem Data , 1998 .

[5]  C. Heipke,et al.  The evaluation of MEOSS airborne three-line scanner imagery : Processing chain and results , 1996 .

[6]  Peter Reinartz,et al.  Comparison of the Accuracy of DEM from SPOT HRS two-fold Stereo Data and HRS/HRG three-fold Stereo Data in Barcelona Test Site , 2004 .

[7]  T. Toutin,et al.  DEM Generation with ASTER Stereo Data , 2001 .

[8]  T. Toutin.  Generation of DSMs from SPOT-5 in-track HRS and across-track HRG stereo data using spatiotriangulation and autocalibration , 2006 .

[9]  M. Lehner,et al.  Semi-Automatic Derivation of Digital Elevation Models from Stereoscopic 3-Line Scanner Data , 1992 .

[10]  Peter Reinartz,et al.  Accuracy for DEM and Orthoimages derived from SPOT HRS Stereo Data without using GCP , 2004 .

[11]  Michael Eineder,et al.  SRTM and beyond: current situation and new developments in spaceborne InSAR , 2003 .

[12]  Rupert Müller,et al.  Verification of digital elevation models from MOMS-2P data , 2001 .

[13]  A. Bouillon,et al.  SPOT 5 HRS geometric performances: Using block adjustment as a key issue to improve quality of DEM generation , 2006 .

[14]  Peter Reinartz,et al.  Comparison and fusion of DEM derived from SPOT-5 HRS and SRTM data and estimation of forest heights , 2005 .

[15]  Ramon Alamús,et al.  DEM generation from SPOT-5 3-fold along track stereoscopic imagery using autocalibration , 2006 .