DATA PRECONDITIONING WITH GABOR NONSTATIONARY DECONVOLUTION FOR RADAR IMAGING OF HIGHLY DISSIPATIVE AND DISPERSIVE MEDIA

In medical microwave imaging applications, electromagnetic (EM) waves propagate through human tissues, which are inherently attenuative and dispersive. In the resulting image, these effects translate to a lack of resolution that increases with time/distance. To produce microwave images with high resolution, there is a strong need for a technique that is able to compensate for the energy loss and correct for the wavelet distortion. Gabor nonstationary deconvolution was developed in the field of Seismology to compensate for attenuation loss, correct phase dispersion, and produce images with high resolution. In this study, the Gabor algorithm is proposed to deal with the nonstationarity in EM wave propagation and attenuation. Gabor deconvolution is essentially based on the assumption that the anelastic attenuation of seismic waves can be described by a constant Q theory. We investigate the Q characterization of EM wave propagation, the frequency-dependency of EM Q, and the effectiveness of Gabor deconvolution to deal with high loss and dispersion. To accommodate for the EM application conditions, several adjustments are made to the proposed algorithm. Our test results indicate that Gabor nonstationary deconvolution is able to sufficiently compensate for attenuation loss and correct phase dispersion for EM waves that propagate through lossy and dispersive media.