Study on the geolocation algorithm of space-borne SAR image

SAR (Synthetic Aperture Radar) images have been widely used nowadays, as the SAR system is capable of scanning the earth surface objects into high resolution images. Before the SAR images are used, the geolocation step is needed to locate arbitrary pixels in an image accurately. Geolocation is very important in geometric rectification, geocoding, as well as object location and detection in the SAR image. In this paper, we propose a novel geolocation algorithm, which is a hybrid of an iterative algorithm and the conventional analytic algorithm based on Range-Doppler (RD) location model. First a new analytic algorithm adopted in our approach is presented. Next, in order to correct the geometry and terrain height, an iterative routine is integrated into the procedure. The experiment results indicate that our algorithm is efficient and can achieve higher accuracies compared with three state-of-the-art location algorithms.