Non-Linear Reconstruction Of Dielectric Target In Dispersive Media Featuring Ultra-Wide Band Sensing And Gradient Minimization
Loading...
Date
2011-06
Authors
Zanoon, Tareq F. A.
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
In the last decade, there has been a growing interest in using UWB signals for
microwave tomography. The diversity of frequencies in the illuminating UWB signal
offers a unique combination enabling scattering of very long to very short wavelengths,
and thus, collecting more information about the target. While microwave in general
posses properties of preference for many imaging applications, inversion algorithms
leading to recovery of the dielectric profile are complex in their nature, and vulnerable to
noisy experimental conditions and environment. In this study, the experimental
feasibility of UWB microwave tomography is investigated for quantitative image
reconstruction. A simplified yet robust gradient based iterative image reconstruction
technique is developed to adapt to the noisy experimental conditions. The technique is
based on the conjugate gradient method with Polak-Ribière’s non-quadratic approach,
while the gradient is obtained using the central difference technique and Broyden’s
updating strategy. The feasibility and accuracy of this new inverse scattering algorithm
is compared with the adjoint method with forward and reverse time-stepping algorithm.
Image reconstruction is performed on synthesized numerical data, and experimental
measured data, obtained via a specifically designed data acquisition system. Results
indicate the super-resolution capability where the detection of millimeter scaled targets
corresponding to one tenth of a wavelength from simulated data, and one quarter of a
wavelength via experimental data were obtained in a relatively complicated structure.
Description
Keywords
Non-linear reconstruction of dielectric target , dispersive media featuring ultra-wide band sensing