Processing video-SAR data with the fast backprojection method

Video synthetic aperture radar (video-SAR) is a land-imaging mode where a sequence of images is continuously formed when the radar platform either flies by or circles the scene. In this paper, the fast backprojection (FBP) algorithm is introduced for video-SAR image formation. It avoids unnecessary duplication of processing for the overlapping parts between consecutive video frames and achieves O(N2 log N) complexity through a recursive procedure. To reduce the processing complexity in video-SAR system, the scene is partitioned into the general region (GR) and the region of interest (ROI). In different regions, different aperture lengths are used. The proposed method allows a direct trade between processing speed and focused quality for the GR, meanwhile reserving particular details in the ROI. The effectiveness is validated both for a simulated scene and for X-band SAR measurements from the Gotcha data set.