Doctor of Philosophy, The Ohio State University, 2023, Electrical and Computer Engineering

Rough surface scattering is an essential aspect of modern remote sensing research, as virtually all real-world surfaces exhibit some degree of roughness whose effects cannot be adequately accounted for using simple planar surfaces. However, knowledge of what each rough surface model is capable of is critical, as choosing an appropriate model will aid in providing accurate results while minimizing the computational cost incurred. To explore these capabilities, four studies were conducted to assess how various rough surface scattering models fare in scenarios of current interest to the remote sensing community. The first two studies involve the Kirchhoff approximation (KA), with the first study assessing its applicability when the normalized coherent reflected field (which the KA is commonly used to model) is -20dB or lower, and the second study comparing it to a second-order correction term based on the second-order small slope approximation (SSA2) for ocean surface scattering. The first study shows that the KA continues to be applicable for such low amplitude cases, and the second study shows that the second-order correction shows no marked improvement over the base KA overall. The third study uses the SSA2 to validate retrieved zero-Doppler delay waveforms as part of a campaign to explore off-specular ocean scattering, and found the model waveforms to match the retrieved waveforms well in most cases considered. The fourth and final study uses simulated SAR imagery to determine under what conditions a monostatic radar system will observe the same surface scattering as a bistatic radar system, and revealed that cases with near-normal incidence angles and minor roughness yield the best agreement, with effects such as shadowing and multiple reflections accounting for most of the disagreements.

Committee: Joel Johnson (Advisor); Fernando Teixeira (Committee Member); Robert Burkholder (Committee Member) Subjects: Electrical Engineering; Electromagnetics; Physics; Remote Sensing

Master of Science (MS), Wright State University, 2018, Physics

The Generalized Harvey-Shack (GHS) surface scatter theory claims to predict the scatter from rough surfaces over a wide domain of validity but has not been widely implemented or tested. The focus of this research is to provide a documented implementation and experimental validation of the GHS theory for the prediction of the bidirectional reflectance distribution function (BRDF) from real rough surfaces. First, the GHS theory was developed for the MATLAB software and the details of its implementation and application to experimental comparison are discussed. Next, a BRDF measurement system was designed, built, and validated. The BRDF of the Spectralon reflectance standard was measured with high agreement. Sample surfaces were characterized via atomic force microscopy as inputs to the GHS theory. The BRDF of each surface was then measured and simulated with the assumption of Gaussian autocovariance. When this assumption held, the GHS theory predictions matched well with experimental results.

Committee: Jason Deibel Ph.D. (Advisor); Brent Foy Ph.D. (Committee Member); Jerry Clark Ph.D. (Committee Member) Subjects: Optics; Physics

Doctor of Philosophy, The Ohio State University, 2020, Electrical and Computer Engineering

Techniques for conducting spaceborne earth remote sensing are well established in the literature. Existing approaches include active and passive methods typically involving the launch of dedicated satellite platforms into orbit. More recently, there has been increasing interest in a relatively less mature mode of remote sensing, Global Navigation Satellite Signal Reflectometry (GNSS-R), which has opened new venues of investigation for the retrievals of geophysical parameters of interest at a global scale with unprecedented spatial and temporal coverage at a fraction of the cost compared to conventional satellite missions. This dissertation aims to support the use of spaceborne GNSS-R observations for global land and ocean remote sensing through investigating the nature and dependencies on surface geophysical properties of these returns and by developing algorithms to retrieve those of interest. The utility of the proposed analyses and methodologies are investigated in the context of NASA's Earth Venture Mission, CYGNSS (Cyclone Global Navigation Satellite System). For studies of land remote sensing, a time series retrieval method is introduced for near surface volumetric soil moisture content retrievals. This is supported by an analysis of the physical dependence of GNSS-R DDMs on land properties, showing that variations with soil moisture and composition, vegetation cover, and surface roughness are all to be expected. The proposed time series retrieval algorithm leverages the slowly varying nature of many of these processes to retrieve soil moisture. Development of complementary approaches to reduce the corrupting effects of low SNR and coherent DDMs in addition to compensating for incidence angle variability are all presented. The utility of the proposed methodology is first explored with simulated GNSS-R measurements and is subsequently extended to CYGNSS measurements on the local, regional and global scales. In support of the broader science community (open full item for complete abstract)

Committee: Joel Johnson (Advisor); Fernando Teixeira (Committee Member); Ethan Kubatko (Committee Member) Subjects: Electrical Engineering; Electromagnetics

Doctor of Philosophy, The Ohio State University, 2019, Electrical and Computer Engineering

Remote sensing of the ocean surface has become an invaluable technique in advancing the scientific understanding of the Earth and its physical and biological processes. Various land-, sea-, air-, and space-based platforms have been deployed over the years with a wide array of instruments ranging from synthetic aperture radar (SAR) imagers to scatterometers for ocean observing missions. This thesis primarily focuses on two such datasets: a high resolution, high revisit rate L-band SAR dataset collected using the SMAP satellite and a coherent-on-receive X-band dataset collected from a wave-sensing radar using a ship-based platform. Investigations of these datasets presented in this thesis are intended to advance the remote sensing and physical modeling capabilities of ocean surfaces. The SMAP 1 km resolution data collected at a global revisit rate of 2-3 days provides a dramatic improvement on temporal and spatial sampling over previous SAR missions. The backscatter normalized radar cross section (NRCS) measurements in both co- and cross-polarizations are used for sea surface modeling studies using empirical and physical models. The backscatter NRCS predictions made using the two-scale model (TSM) and the second-order small slope approximation (SSA2) under assumed wind-only fully-developed sea surface conditions indicate the presence of swell waves. A wind + swell combined model of the ocean surface is then used to characterize the impact due to swell waves on SMAP NRCS measurements. Results indicate a substantial improvement on backscatter NRCS prediction capability using the combined model over the wind-only model results. The SMAP dataset is further investigated for swell features, while the swell prediction capability is also assessed. A team from The Ohio State University and the University of Michigan developed a low-cost coherent-on-receive X-band wave sensing radar using commercial off-the-shelf (COTS) components for ship-based applications. The radar (open full item for complete abstract)

Committee: Joel Johnson (Advisor); Robert Burkholder (Committee Member); Graeme Smith (Committee Member); Caglar Yardim (Committee Member) Subjects: Electrical Engineering; Electromagnetics

Doctor of Philosophy, The Ohio State University, 2007, Electrical Engineering

The Small Perturbation Method (SPM) is a low frequency approximation to the electromagnetic scattering from rough surfaces. The theory involves a small height expansion in conjunction with a perturbation series expansion of the unknown scattering coefficients. Recently, an arbitrary order, iterative solution procedure has been derived for SPM: kernels at any order are expressed as a summation over lower order kernels in an iterative fashion. Such a form is very useful, because it allows evaluation of the field statistical moments in a direct manner, when considering stochastic surfaces. In this dissertation, this procedure is extended to the two layer (two rough surfaces on top of each other) problem and the complete solution is given. Utilizing this formulation, the second and fourth order bi-static scattering coefficients for two rough surfaces characterized by two uncorrelated Gaussian Random Processes (GRP) are obtained. The effects of upper and lower roughnessesand the interaction effect in the total fourth order cross section can be identified in the theory. Studies on the ratio of the interaction effect to the total cross section are presented for example cases, investigating the relative importance of interactions among surfaces. Results show the interaction term contributes most to the cross-pol cross sections when surfaces are close to each other at near grazing incidence. In addition, the previously developed arbitrary order SPM solution for the single layer problem is utilized to derive the fourth order term in the small slope approximation (SSA) of thermal emission from the sea surface. It is shown that this term has the form of a four-fold integration over a product of two sea spectra for a Gaussian random process sea, thereby describing emission “interaction” effects among pairs of sea waves. Interaction effects between “long” and “short” waves are considered, both through numerical and approximate evaluations of the fourth order theory. The approxima (open full item for complete abstract)

Committee: Joel Johnson (Advisor) Subjects: