Please use this identifier to cite or link to this item:
|Title:||Inverse Solutions in Electromagnetism with Applications in Biomedical Imaging and Non-Destructive Testing|
|Authors:||Amineh, Reza K.|
|Advisor:||Nikolova, Natalia K.|
Reilly, James P.
|Department:||Electrical and Computer Engineering|
|Keywords:||electrical and computer engineering;inverse solutions;electromagnetism; microwave engineering;biomedical imaging;non-destructive testing;magnetic flux leakage|
|Abstract:||<p> This thesis presents solutions to several inverse problems m electromagnetism and microwave engineering. In general, these inverse problems belong to two applications: breast cancer diagnosis using microwave imaging and defect characterization in metallic structures using magnetic flux leakage (MFL). </p> <p> Our contribution in microwave imaging for breast tumor detection can be divided into three parts. First, we propose a novel ultra-wide band (UWB) antenna that can operate in direct contact with the breast without the need for coupling liquids. This antenna is designed such that more than 90% of the radiated power is directed toward the tissue from its front aperture over the UWB. The performance of the antenna is investigated via simulation and measurement of the following parameters: return loss, near-field directivity, efficiency, fidelity, and group velocity. Overall, the results show that the antenna is a good candidate for frequency and time-domain imaging techniques. </p> <p> Second, we have proposed an aperture raster scanning setup that benefits from the features of our novel antenna. In this scanning setup, the breast tissue is compressed between two rectangular plates (apertures) while two antennas perform two-dimensional (2-D) scan by moving together on both sides of the compressed tissue. For each scanning step, the transmission S-parameter between the two antennas is recorded at several frequencies within UWB. Then, the modulus of the calibrated transmission S-parameter is plotted at each frequency to provide a 2-D image of the interior of the breast. The images are enhanced using a de-blurring technique based on blind de-convolution. This setup provides real time images of strong scatterers inside the normal tissue. </p> <p> Third, we propose 2-D and three-dimensional (3-D) holography algorithms to further improve the quality of the images obtained from the proposed planar scanning setup. These techniques are based on the Fourier transforms of the collected data to provide an image of a 2-D target (when collecting data at a single frequency) or a 3-D target (when collecting wide-band data). These techniques are fast and very robust to noise. The capability of the 2D and 3-D holographic imaging techniques is examined via simulation results. </p> <p> For defect characterization in metallic structures using MFL technique, we propose fast and reliable methodologies to invert the measured MFL response to the defect's shape parameters. First, we present a procedure to estimate the shape parameters of rectangular cracks which are the most common type of defects in the metallic structures. The procedure consists of estimating orientation, length, and depth of the cracks, consecutively. We validate this procedure via estimating the shape parameters of pre-known cracks from the simulated and measured MFL responses. Then, we present a methodology based on space mapping (SM) optimization for defect characterization. We examme the efficiency of this methodology for two types of defects: rectangular cracks and cylindrical pits. </p>|
|Appears in Collections:||Open Access Dissertations and Theses|
Files in This Item:
|Amineh_Reza_K_2010_PhD.pdf||50.68 MB||Adobe PDF||View/Open|
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.