Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Kai_Dissertation.pdf (8.18 MB)

ETD Abstract Container

Abstract Header

Physics-Based Near-Field Microwave Imaging Algorithms for Dense Layered Media

Ren, Kai
http://orcid.org/0000-0003-4704-5743
http://rave.ohiolink.edu/etdc/view?acc_num=osu1511273574098455

Abstract Details

2017, Doctor of Philosophy, Ohio State University, Electrical and Computer Engineering.
It is of importance to understand the physics as electromagnetic (EM) waves propagate through stratified media, are scattered back from buried irregularities, and are received by an antenna in the near field. To generate better images, we need to incorporate the physics of the phenomena into the imaging algorithm, such as multiple reflections, refractions resulting from the existence of interfaces, and diffractions from embedded targets. A forward model is developed based on the spectral Green’s function associated with layered media weighted by the antenna gain pattern, satisfying the near-field condition and incorporating all refraction effects. Thereby, the weak scattering from deeper layers and wide angles will be compensated in a model-based imaging algorithm with the consideration of the refraction coefficients and gain pattern, respectively. To form real-time continuous images of targets embedded in a layered structure, a near-field uniform diffraction tomographic (UDT) imaging algorithm is developed. Conventional diffraction tomography (DT) improperly applies the stationary phase method for stratified environments to evaluate the innermost spectral integral. In DT the large argument is assumed to be the depth, which is not appropriate for near-field imaging. This results in amplitude discontinuities occurring at the interfaces between adjacent layers. The correct dimensionless large argument is the product of the free space wavenumber and the depth, as used in high-frequency asymptotic solutions. UDT therefore yields uniformly continuous images across the interfaces. And like DT, UDT retains the fast Fourier transform (FFT) relation in the algorithm for generating images very efficiently. Both 2D and 3D cases are investigated to verify the efficacy of the proposed UDT algorithm. To overcome the singularity problem caused by nulls in the antenna gain pattern in DT and UDT, a fast back-projection (FBP) imaging algorithm is propose to provide balanced monostatic and bistatic images, where both the stationary phase method and FFT are implemented to achieve the same computational efficiency as DT and UDT. FBP is derived based on the conventional back-projection (BP) imaging algorithm, and then finding the Fourier transform relation after applying the matched filter to the forward model. The comparison between UDT and FBP is investigated for the free space case. FBP is also demonstrated by a combined monostatic/bistatic synthetic aperture radar (SAR) imaging system, where the time for data collection is reduced by half through the appearance of virtual array elements. To achieve even faster data collection, a fixed multi-static array of antennas is proposed to efficiently illuminate a given subsurface volume using a simultaneous multiple-input multiple-output (MIMO) type signal. After constructing a radar image, it is of interest to quantify buried target parameters, such as the shape, location, orientation, and size. These parameters can be estimated by the implementation of the spatial moment method. However, due to the lens effect when EM waves propagate through a slab with a high dielectric constant, embedded objects which are roughly symmetric such as spheres can look similar in size regardless of the actual size. To help overcome this issue, another unique feature of images in layered media is employed, namely, the presence of shadows. Due to the lens effect, the projection of an embedded object onto the next deeper interface tends to show better size information than its primary image. However, the shadow image can be fuzzy if a conventional imaging algorithm is applied, decreasing the accuracy of size estimations. Modification using a local spatial filter for improving shadow images is investigated. Lastly, it has been shown that the interface between the external region and a dense medium tends to focus the energy from the antenna straight into the medium due to refraction, losing important wide-angle illumination of embedded objects. This phenomenon is investigated further to exploit the Brewster angle effect in order to launch energy at wide angles. Illumination with a dipole oriented normal to the layered medium is simulated because it has a null pointing straight into the medium and strong radiation into the Brewster angle direction. When combined with a transverse source, it is shown that the illumination of targets buried in a dense dielectric media is much improved.
Robert Burkholder, Dr. (Committee Member)
Fernando Teixeira, Dr. (Committee Member)
Graeme Smith, Dr. (Committee Member)
188 p.
Electrical Engineering; Electromagnetics
Microwave imaging, Radar signal processing, Uniform diffraction diffraction, fast back projection, least squares method, imaging through periodic structure, spatial moments method, buried object size estimation, shadow projection, lens effect

Recommended Citations

Citations

  • Ren, K. (2017). Physics-Based Near-Field Microwave Imaging Algorithms for Dense Layered Media [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1511273574098455

    APA Style (7th edition)

  • Ren, Kai. Physics-Based Near-Field Microwave Imaging Algorithms for Dense Layered Media. 2017. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1511273574098455.

    MLA Style (8th edition)

  • Ren, Kai. "Physics-Based Near-Field Microwave Imaging Algorithms for Dense Layered Media." Doctoral dissertation, Ohio State University, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=osu1511273574098455

    Chicago Manual of Style (17th edition)

osu1511273574098455
977
© 2017, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.