Abstract
Conference Title: 2015 Signal Processing Symposium (SPSympo) Conference Start Date: 2015, June 10 Conference End Date: 2015, June 12 Conference Location: Debe, Poland Multistatic SAR/ISAR can be described and generalized using the principles of radar (or radio frequency) tomography. In radar tomography, distributed transmitters and receivers sense a region of interest using suitable waveforms. Using the principles of linear scattering (Born approximation), a (linear) relation exists between the measured returns and the shape of targets, and an image can be formed by inverting such relation. Due to the limited illumination and observation points, each point target will exhibit a spatially-varying point spread function (psf). A dominant scatterer having a large psf will inevitably mask the return of weak scatterers nearby. To mitigate this masking effect, the image is analyzed to identify dominant scatters. Then, these scatterers are modelled as dipole sources and included in the imaging formation as part of illumination points (i.e., new transmitters). When dominant scatterers are modelled as transmitters, their respective sidelobes are removed from the image, so that weak targets can be identified. Simulations and results demonstrate this concept.