| The spotlight mode SAR provides the opportunity to collect SAR data over a wide azimuth angular interval (independent of antenna azimuth beamwidth) with the potential for extremely fine azimuth resolution. However, the MTRC soars severely with the increasing of the azimuth width and resolution. Without compensation, MTRC prevents the realization of fine resolution over a reasonable scene size. Therefore, the demand of imagery with both fine spatial resolution and large processed scene size, which is the orientation of the dissertation, presents difficult challenges to SAR image formation algorithms.In this dissertation, the common choices of spotlight mode SAR image formation, such as PFA and RMA, as well as new emerging methods like WPFA and DDA, are discussed in a heuristic and comprehensive way, provided with the simulation illustrations and real data proccessing results. Each algorithm processes unique capabilities and limitations that determine its applicability for specific SAR systems and missions. PFA suffers from range curvature due to residual MTRC. Geometric distortion is traditionally corrected by an image resampling operation which is applied as a post-processing step to PFA. The scene size limits of PFA presented by range curvature QPE, can be alleviated by using an efficient image domain space-vairant post-filter. In the meanwhile, RMA, WPFA and DDA are potentially ideal choices for fine-resolution SAR imaging of large scene in challenging environments that include imaging at a close range, using low centre frequency, and imaging very large scenes. However, their computational burden is also considerable.Efficiency is another question. Some improvement is made in this dissertation based on compare and conclusion of existing methods and theories, such as substituting space-variant filter for range interpolation in geometric distortion correction and improving ATAFS, both of which are demonstrated to reduce computational burden effectively. |
PDF Full Text Request