Model-based Correction of Through-wall SAR Imagery via Raytracing

This paper develops a polarimetric model for through-wall propagation effects at RF frequencies and incorporates this model in a backprojection imaging framework that corrects for defocusing observed with standard methods. This model accounts for refractive phase delay as well as multiple internal reflections within the wall, and the image-reconstruction correction refocuses the entire scene behind the wall at once. The proposed algorithm that implements this correction leverages CAD raytracing and forward prediction capabilities from widely-used shooting-bouncing-rays (SBR) electromagnetic scattering tools, enabling efficient implementation on complex wall geometry (e.g., windows, doors, and multiple walls). The development in this paper suggests that fully-polarimetric data is required in general to recover the focused reflectivity of the scene behind the wall for any arbitrary transmit and receive polarizations; however, for scenes without significant vertical extent the proposed algorithm can refocus a single-polarization measurement. This paper presents a benchmark scene comprising a homogeneous dielectric wall and four trihedral reflectors, two of which are obstructed by the wall. This paper uses synthetic data of the scene, generated using a nearfield SBR implementation, to demonstrate the effectiveness of the proposed algorithm.