Target aspect estimation from single and multi-pass SAR images

A technique is presented for estimating the aspect of targets in SAR imagery for use in indexing, feature extraction and recognition. Aspect estimation is enhanced by combining multiple images of the same target. In order to properly combine the estimation of multiple passes, it is necessary to accurately register the images to a common coordinate frame. An algorithm for registering multiple high resolution SAR images, is presented. A global affine transformation derived from the sensor acquisition parameters is used to automatically register the images, followed by a refinement to correct for translational errors. The registered SAR images are used for improving the estimates of target orientation angles, detecting the presence of occlusion and indicating poor target segmentation.