SYNTHETIC APERTURE RADAR IMAGING AND POST-IMAGING PROCESSING

Abstract

SAR is a high resolution imaging radar used to generate maps of surface target areas and terrains. It is different from most other remote sensing techniques in that the received raw data must be coherently processed to produce an image. In this project, Range-Doppler algorithm is developed to obtain SAR image from ESAR raw data. The conditions of the algorithm, influence of linear frequency modulation (LFM) chirp rate and weighting on system resolution and sidelobe are investigated. Furthermore, the formation of speckle in SAR imaging process and its reduction techniques both in spatial domain and frequency domain have been studied. Images of the coastal area obtained by SAR are often contaminated by speckle as well as other strong noise returns caused by winds and waves, with amplitudes fluctuating with the incident angle of the radar beam. A novel algorithm for SAR image enhancement and feature extraction is proposed in the post-imaging process. This algorithm is different from the conventional methods in that it needs no a priori information about the relative roughness of the land and sea. In this algorithm, a seed search method is used to classify the region instead of the common edge tracing algorithms. Finally, a pseudo-colored image of Singapore using ERS-1 data is obtained, showing the effectiveness of the algorithm.