Spatial accuracy analysis of Sentinel-1 SAR satellite images by comparing maps

Abstract

Synthetic Aperture Radar (SAR) images are used in several different applications for Remote Sensing purposes. SAR is an imaging sensor that can detect high-resolution ground images under a wide variety of imaging conditions. As SAR is an active system, the data are already acquired with geo-position information. To investigate to verify the image spatial accuracy, a part of the Antalya region of Turkey was selected as test site. The Open Street Map (OSM) and Photogrammetric Digital Map (PDM) data of the area on the Antalya O25 map were used in comparison. First, characteristic common points were selected on the OSM data and the SAR satellite image both. The projected coordinates of these points were calculated with the QGIS software. Normal distribution of the coordinate differences in these data sets were plotted. It was confirmed that the data sets were in normal distribution and standard deviation and 2 * standard deviation values were calculated. The maximum and minimum confidence interval (95%) was determined according to the standard deviation limit values. X and Y coordinate differences were calculated for 49 selected points from both image pairs SAR&OSM and SAR&PDM. Finally, the maximum differences show that the SAR positional accuracy respect to OSM and PSM is below 1 pixel azimuthal resolution.