DBPapers
DOI: 10.5593/SGEM2016/B22/S10.117

INTEGRATION OF DIGITAL SURFACE MODELS FROM PHOTOGRAMMETRIC AND AIRBORNE LASER SCANNING DATA FOR THE PROCESSING OF RADAR IMAGERY

B. Kraszewski, A. Karwel, Z. Kurczynski, D. Ziolkowski
Thursday 8 September 2016 by Libadmin2016

References: 16th International Multidisciplinary Scientific GeoConference SGEM 2016, www.sgem.org, SGEM2016 Conference Proceedings, ISBN 978-619-7105-59-9 / ISSN 1314-2704, June 28 - July 6, 2016, Book2 Vol. 2, 917-924 pp

ABSTRACT
Data from photogrammetry measurements and airborne laser scanning are excellent source for accurate Digital Surface Models (DSM) development. These models are used for various purposes, including deformation studies with Persistent Scatterer Interferometry with the use of high resolution radar images. The Polish geospatial resources have large databases of these models obtained within lidar and photogrammetric projects. The best accurate models for high resolution image processing are generated from LiDAR data but unfortunately these data don’t cover the whole country. For areas for which we don’t have overlapping height models we can apply existing lidar elevation data. We can use to it elevation data derived from aerial image stereo matching. This paper presents the method of generating DSM from integration of different source data for radar image processing. It is very important task, as we need DSM of comparable accuracy for the whole radar scene to get consistent interferometric results. Output DSM had 3 meter resolution as Cosmo-SkyMed high resolution radar images. To generate this model we used DSM from aerial images matching and generalized DSM from models with 0.5 m resolution from LiDAR data.
Additional DSM from image matching has been subjected to manual errors improvement. In the accuracy analysis of DSM we compare heights for both integrated models for different land cover areas. Average difference between models height was - 30 cm for all analyzed types of terrain, and standard deviation was 4 m. In case of flat area difference was -14 cm and standard deviation was 35 cm. The conclusion from integration is that in flat and urban areas we can integrate height models from LiDAR and photogrammetry matching to radar data processing.

Keywords: digital surface model, stereomatching, aerial images, LiDAR, integration