A new approach to distributed passive radar imaging by 2-D NUFFT

This paper studies distributed passive radar imaging of two-dimensional (2-D) scene. The echoes and the reflectivity of target are confirmed to be a Fourier transform pair. However, due to the space configuration based on the illuminators of opportunity, the echoes are sampled non-uniformly, which lead to the non-equispaced property of space-spectrum distribution. In this case, direct 2-D FFT reconstruction doesn´t apply now. To retain the fast computing of image reconstruction like FFT, non-uniform fast Fourier transform (NUFFT) is introduced to avoid complex interpolation. We generalize NUFFT to 2-D NUFFT to tackle the problem that the space-spectrum is unequally-spaced in both directions of μ and v. This new 2-D NUFFT method is compared with polar-coordinate method in efficiency and accuracy using both the analytical and numerical approaches. Simulations demonstrate that 2-D NUFFT can reduce the computational complexity while ensuring image reconstruction accuracy. So it´s a promising strategy for realtime imaging of distributed passive radar.