Height Retrieval in Postprocessing-Based VideoSAR Image Sequence Using Shadow Information

Video synthetic aperture radar (VideoSAR) imaging mode presents the capability of generating dynamic scenes. In this paper, a postprocessing-based VideoSAR imaging scheme suitable for frequency-domain SAR algorithms is proposed to achieve more flexible application. First, the full-aperture image can be partitioned into the stationary region and a dynamic region of interest (DROI). Then the selected DROI data is segmented at a certain overlapping rate in the azimuth Doppler spectrum. Thus, the unified VideoSAR image coordinate system is achieved, and hence extra image registration technique is avoided. Moreover, VideoSAR can distinguish the relationship between the stationary target and shadow for a more accurate retrieval of height information. Subsequently, a complete procedure of information extraction without geometric
prior
knowledge, including denoising algorithm, threshold segmentation, matching strategy, and height retrieval is proposed in detail. Based on varying shadows and scattering characteristics in the DROI, multiple experiments using real airborne X-band VideoSAR data validate the feasibility of the proposed imaging scheme and the height retrieval method can achieve a higher accuracy.