SAR Based Non-Destructive Evaluation of Irregularly Shaped Objects with Simultaneous Estimation of Geometry and Permittivity

The geometrically correct registration of subsurface structures that are embedded in a medium with unknown shape and permittivity - e.g. a small defect in fiber or ceramic components - is one of the most challenging tasks in radar imaging. This paper presents a new methodology for subsurface synthetic aperture radar (SAR) imaging with integrated estimation of geometry and permittivity applicable for nondestructive evaluation (NDE). Two SAR measurements are performed, one scanning the device under test (DUT) from the front side, the second scanning the DUT from the backside. By comparing the contour profile derived from the front side/ backside measurement with an SAR image of the complete DUT reconstructed under the assumption of free-space propagation, the permittivity and complete contour of the DUT are estimated. A high-resolution image is then reconstructed with an adapted subsurface backprojection algorithm that considers the contour and permittivity estimates correctly. Experimental results with a polyethylene DUT with two small embedded defects are presented that validate the proposed technique. It is shown that with the novel approach good estimations of shape and permittivity as well as geometrically correct high resolution images can be achieved.