Optimum 2-D Nonuniform Spatial Sampling for Microwave SAR-Based NDE Imaging Systems

Microwave and millimeter-wave synthetic aperture radar (SAR)-based imaging techniques, which are used for nondestructive evaluation (NDE), have shown tremendous usefulness for the inspection of a wide variety of complex composite materials and structures. An important practical issue associated with these imaging techniques is the required criteria associated with the physical gathering of the imaging data. In previous work on uniform sampling optimization, it was shown that the (uniform) spatial sampling density should be higher than the Nyquist density if preservation of spatial resolution, as defined by the half-power width of an ideal point target, is of interest. Conversely, nonuniform sampling has shown to provide effective signal reconstruction even for average spatial sampling densities below the Nyquist density-a distinct advantage over uniform sampling. This paper presents a comprehensive study into the optimization of nonuniform sampling for microwave SAR-based NDE imaging using three typical reconstruction techniques for nonuniformly sampled data, including natural interpolation, area-weighted Fourier integration, and conjugate gradient residual error minimization methods. To study the efficacy of these reconstruction methods, simulations of a point target were performed for a range of target distances and a range of average spatial sample separations. Resulting SAR images from the reconstruction techniques are then analyzed and compared according to two metrics: error between an image and an ideal image and resolution as determined by the half-power width of a point target. This is followed by experimental results corroborating the simulation results. Finally, nonuniform sampling requirements, given a minimum metric performance, are generalized for a given imager aperture size.