An efficient method for three-dimensional through-the-wall radar imaging

Abstract In three-dimensional (3D) through-the-wall radar imaging, unknown wall parameters such as thickness and relative permittivity distort electromagnetic wave propagation, causing image defocusing and target displacement. Autofocusing is therefore essential, but conventional methods are computationally demanding due to two-dimensional (2D) searches over wall parameters and repeated reconstruction of full 3D images. To address this challenge efficiently, a slice-based 3D imaging strategy is adopted. Radar data are acquired along a linear aperture parallel to the wall at multiple heights. For each height, the collected measurements form a B-scan dataset, which is processed independently to reconstruct a 2D image slice. A computationally efficient two-step framework is proposed. First, the wall’s relative permittivity is estimated from the time-domain response of the B-scan at a selected height, and autofocusing is performed by varying only the wall thickness, reducing the conventional 2D search to a one-dimensional problem. Second, focused 2D image slices are reconstructed at all heights. The final 3D image of the concealed targets is obtained by superimposing these slices into a volumetric representation. The experimental results demonstrate that the proposed technique yields well-focused images while significantly reducing the computational cost.