Navigation with SAR and 3D-map aiding

This paper presents a method for matching spotlight Synthetic Aperture Radar (SAR) images with a georeferenced 3D-map as means for navigational aid. A hypothesis of the flying platform´s absolute position, velocity and direction - which later can be used to correct the inertial navigation system - is attained by image matching and optimization. A projective model with 6 DoF is used to create a simulated SAR image from a 3D map. The parameters of the projective model represents the most important of the platform´s navigation state, and these are adjusted by Chamfer matching the captured SAR image to simulated ones. The performance is demonstrated on real spotlight SAR images and 3D-map, and the error is shown to be only a few pixels in average, which in our case is about 3 meters.