Application of FFT and DFT for image reconstruction of planar arrays in microwave holographic diagnostics

It is demonstrated that, for a planar array of discrete elements, the accuracy of the reconstruction is more dependent upon the location of the reconstruction points than on the actual resolution (i.e., spacing) of the reconstruction points. The surface field distribution of a planar array of infinitesimal dipoles is reconstructed from complex far-field and near-field data with the use of microwave holographic metrology. A comparison is made between the discrete Fourier transform (DFT) and the more traditional fast Fourier transform (FFT) to reconstruct the surface distribution. The former technique gives the user more freedom in choosing the location of the reconstruction points. A 22 by 22 element planar array of infinitesimal dipoles was simulated. The FFT required less computer time to perform the reconstruction than the DFT, but the increased accuracy provided by the DFT warrants its use in cases where the element anomalies are small or a more accurate determination of the amplitude and phase needs to be made.<>