Search

Ground-Penetrating Radar (GPR) is a popular non-destructive electromagnetic (EM) technique that is used in diverse applications across different fields, most commonly geophysics and civil engineering. One of the most common applications of GPR is concrete scanning, where it is used to detect structural elements and support the assessment of its condition. However, in any GPR application, the data have no resemblance to the characteristics of targets of interest and a means of extracting information from the data regarding the targets is required. Interpreting the GPR data, to infer key properties of the subsurface and to locate the targets is a difficult and challenging task and is highly dependent on the processing of the data and the experience of the user. Traditional processing techniques have some drawbacks, which can lead to misinterpretations of the data in addition to the interpretation being subjective to the user. Machine learning (ML) has proven its ability to solve a variety of problems and map complex relationships and in recent years, is becoming an increasingly attractive option for solving GPR and other EM problems regarding processing and interpretation. Numerical modelling has been extensively used to understand the EM wave propagation and assist in the interpretation of GPR responses. If ML is combined with numerical modelling, efficient solutions to GPR problems can be acquired. This research focuses on developing a numerical equivalent of a commercial GPR transducer and utilising this model to produce realistic synthetic training data sets for deep learning applications. The numerical model is based on the high-frequency 2000 MHz "palm" antenna from Geophysical Survey Systems, Inc. (GSSI). This GPR system is mainly used for concrete scanning, where the targets are located close to the surface. Unknown antenna parameters were found using global optimisation by minimising the mismatch between synthetic and real responses. A very good match was achieved, demonstrating that the model can accurately replicate the behaviour of the real antenna which was further validated using a number of laboratory experiments. Real data were acquired using the GSSI transducer over a sandbox and reinforced concrete slabs and the same scenarios were replicated in the simulations using the antenna model, showing excellent agreement. The developed antenna model was used to generate synthetic data, which are similar to the true data, for two deep learning applications, trained entirely using synthetic data. The first deep learning application suggested in the present thesis is background response and properties prediction. Two coupled neural networks are trained to predict the background response given as input total GPR responses, perform background removal and subsequently use the predicted background response to predict its dielectric properties. The suggested scheme not only performs the background removal processing step, but also enables the velocity calculation of the EM wave propagating in a medium using the predicted permittivity value. The ML algorithm is evaluated using a number of synthetic and measured data demonstrating its efficiency and higher accuracy compared to traditional methods. Predicting a permittivity value per A-scan included in a B-scan results in a permittivity distribution, which is used along with background removal to perform reverse-time migration (RTM). The proposed RTM scheme proved to be superior when compared with the commonly used RTM schemes. The second application was a deep learning-based forward solver, which is used as part of a full-waveform inversion (FWI) framework. A neural network is trained to predict entire B-scans given certain model parameters as input for reinforced concrete slab scenarios. The network makes predictions in real time, reducing by orders of magnitude the computational time of FWI, which is usually coupled with an FDTD forward solver. Therefore, making FWI applicable to commercial computers without the need of high-performance computing (HPC). The results clearly illustrate that ML schemes can be implemented to solve GPR problems and highlight the importance of having a digital representation of a real transducer in the simulations.

Time-zero adjustment or the true ground surface for Ground Penetrating Radar (GPR) applications is a very important aspect and an essential factor in order to carry out accurate shallow depth measurements. As the transmitted and received signals from GPR antennas are affected by the presence of different materials with various dielectric constants and electromagnetic properties adjusting the time-zero appropriately is important. This study uses a realistic Three Dimensional (3D) numerical model of a GPR transducer in order to examine where is the best location for time-zero on a GPR trace. It is shown that in order to establish a robust and consistent time-zero position careful consideration is needed also of the way the two-way travel time of the reflected GPR wavelet is estimated as well. Starting with a simple homogeneous model with a set of different targets a better process of time-zero adjustment and time picking of the GPR wavelets is put forward that is verified using further more complex and realistic heterogeneous models. Further verification is obtained by using experimental data. Estimating the permittivity of heterogeneous mixtures based on the permittivity of their individual components is of high importance with many applications in GPR and in electrodynamics-based sensing in general. The Complex Refractive Index Model (CRIM) is the most mainstream approach for estimating the bulk permittivity of heterogeneous materials and has widely been applied for GPR applications. The popularity of CRIM is primarily based on its simplicity while its accuracy has never been rigorously tested. In the current study, an optimized shape factor is derived that is fine-tuned for modelling the dielectric properties of concrete. The bulk permittivity of concrete is expressed with respect to its components i.e, aggregate particles, cement particles, air-void and volumetric water fraction. Different combinations of the above materials are accurately modelled using the Finite-Difference Time-Domain (FDTD) method. The numerically estimated bulk permittivity is then used to fine-tune the shape factor of the CRIM model. Then, using laboratory measurements it is shown that the revised CRIM model over-performs the default shape factor and provides with more accurate estimations of the bulk permittivity of concrete. Numerical modelling of a heterogeneous concrete model and a bowtie antenna with a separate transmitter and receiver that are able to move independently are also presented in this study. Both models are used for the optimisation of the time-zero position and the CRIM model shape factor.

Within the archaeological record 'mounds' are often ubiquitous. They are common in many ancient cultures, and they vary in size, construction techniques and use. This research is focused upon optimizing the use of remote sensing for the non-invasive study of mounds both in the United States and the United Kingdom. This thesis presents three representative earthen mound sites and proposes a comprehensive and modular survey methodology to guide the planning and execution of a mound survey tailored to the unique requirements presented by the cultural resource at a particular location. In doing so, the research has provided optimized approaches to high resolution three-dimensional topographic models using a variety of digital methods. These models have been shown to accurately capture the variability of the modern ground surface, which is of vital importance to the management of the mounds. Furthermore, these models have proved vital for integrating geophysical methods into the holistic workspace, thereby providing a better archaeological understanding of the below ground remains. Every mound surveyed presented different challenges, and therefore had to be approached in a slightly different way. However, the general methodology was highly effective for both characterizing below-ground archaeological and natural anomalies, and for assessing the state of preservation of all mounds surveyed. As a result, a flowchart has been generated for non-invasive assessment of mounds in general. If followed, this will allow the production of a "snapshot" of the mound or mound group at a fixed point in time with the resolution necessary to produce useful and insightful interpretation. While this research focuses on the application of geophysical and topographic survey in the United Kingdom and United States to a mound or mound group, this methodology and the associated outcomes can be valuable more globally not only for archaeology, but also heritage management.

Humanitarian clearance of minefields is most often carried out by hand, conventionally using a a metal detector and a probe. Detection is a very slow process, as every piece of detected metal must treated as if it were a landmine and carefully probed and excavated, while many of them are not. The process can be safely sped up by use of Ground-Penetrating Radar (GPR) to image the subsurface, to verify metal detection results and safely ignore any objects which could not possibly be a landmine. In this thesis, we explore the possibility of using Full Wave Inversion (FWI) to improve GPR imaging for landmine detection. Posing the imaging task as FWI means solving the large-scale, non-linear and ill-posed optimisation problem of determining the physical parameters of the subsurface (such as electrical permittivity) which would best reproduce the data. This thesis begins by giving an overview of all the mathematical and implementational aspects of FWI, so as to provide an informative text for both mathematicians (perhaps already familiar with other inverse problems) wanting to contribute to the mine detection problem, as well as a wider engineering audience (perhaps already working on GPR or mine detection) interested in the mathematical study of inverse problems and FWI.We present the first numerical 3D FWI results for GPR, and consider only surface measurements from small-scale arrays as these are suitable for our application. The FWI problem requires an accurate forward model to simulate GPR data, for which we use a hybrid finite-element boundary-integral solver utilising first order curl-conforming N\'{e}d\'{e}lec (edge) elements. We present a novel `line search' type algorithm which prioritises inversion of some target parameters in a region of interest (ROI), with the update outside of the area defined implicitly as a function of the target parameters. This is particularly applicable to the mine detection problem, in which we wish to know more about some detected metallic objects, but are not interested in the surrounding medium. We may need to resolve the surrounding area though, in order to account for the target being obscured and multiple scattering in a highly cluttered subsurface. We focus particularly on spatial sensitivity of the inverse problem, using both a singular value decomposition to analyse the Jacobian matrix, as well as an asymptotic expansion involving polarization tensors describing the perturbation of electric field due to small objects. The latter allows us to extend the current theory of sensitivity in for acoustic FWI, based on the Born approximation, to better understand how polarization plays a role in the 3D electromagnetic inverse problem. Based on this asymptotic approximation, we derive a novel approximation to the diagonals of the Hessian matrix which can be used to pre-condition the GPR FWI problem.

The presence and configuration of subglacial water and sediment maintain the fast flow of arterial ice streams in Antarctica and airborne ice-penetrating radar data represent a potential resource of information about the ice-bed interface. In this thesis an original contribution to the exploration of Antarctic subglacial environments is made through the analysis of airborne surveys from Evans, Institute and Möller Ice Streams, West Antarctica. The primary approach employed is the derivation of bed-returned power (BRP), a proxy for ice-bed reflectivity, which is strongly influenced by the presence of liquid water. Estimating radar englacial attenuation (EA) accurately is a critical part of BRP analysis and a modelled approach is primarily used. BRP is derived across Evans Ice Stream and shows large-scale patterns relating to hypothesised hydrological and geological contrasts at the ice-bed. These results are developed to investigate the influence of: (1) adopted EA correction; (2) the influence of assigned ice dielectric properties in modelled EA; (3) subglacial roughness and (4) the spatial scale over which BRP is derived. Some areas of high basal drag can be detected with BRP analysis, indicating that variations in subglacial hydrology are responsible for their existence. The widely-used empirical method of estimating EA by relating ice thickness to uncorrected BRP is shown to be unreliable where ice properties change along a transect. Monte Carlo error analysis of modelled EA shows that poorly constrained ice dielectric properties also result in significant BRP uncertainty. BRP beneath Institute and Möller Ice Streams is derived on catchment- and local-scales over hypothesised subglacial features. Bungenstock Ice Rise is marked with a clear BRP signal but the locations of "active" lakes, as delineated by satellite altimetry, do not. The sensitivity of idealised flow paths to surface change and grid size are investigated. Potential future research directions regarding BRP analysis are discussed.

The “green agenda” combined with highway congestion has accelerated the demand for increased freight and passenger travel on the world’s railways. These increases have driven demand for more efficient and rapid investigation of trackbed ballast. Network Rail and other rail infrastructure operators spend significant financial sums on inspecting, tamping, adjusting, cleaning, and replacing trackbed ballast. Such maintenance is often to the detriment of normal network operation. Industry requires a method of ballast evaluation that is non-intrusive, cheap, can appraise long stretches of track in a short period of time, and give a fingerprinting result from which time-to-maintenance can be calculated and planned. Thus, the aim was to develop evaluation methods using non-destructive testing techniques. A 10-year old full-scale trackbed composed of variously fouled ballast was re-visited and used for experimentation. The condition of the ballast was calculated using the Ionescu Fouling Index. Earlier research at the University of Edinburgh enabled researchers worldwide to characterise ballast using ground penetrating radar (GPR). This research was repeated, validated and taken forward in a series of GPR experiments on the trackbed using a range of antennas from 500MHz to 2.6GHz. New "scatter" metrics were developed to determine ballast condition from the GPR waveforms. These metrics were then used to predict the Ionescu Fouling Index with a correlation coefficient greater than 0.9. One of the current approaches to evaluating the stiffness of railway ballast is to use a Falling Weight Deflectometer (FWD). The viability of using a Prima 100 mini-FWD on railways to measure stiffness was determined and deemed to be ineffective on ballast. The applicability of the impulse response technique on railways was determined. An instrumented hammer was used to excite the ballast, with a geophone measuring the response. The Frequency Response Function of this was successfully correlated with the Ionescu Fouling Index with a correlation coefficient also greater than 0.9. Finally, using GPR data and measured stiffness data collected by Banverket, Sweden, a numerical model to successfully relate radar responses to stiffness was developed.

The early 18th century historic fortification system that surrounded the urban town of Charleston has been a focus of excavations and historical research within the past two decades. However, while the outline of early Charleston’s walled city is hypothesized through the analysis of historic maps and plats, very few fortification locations have been visually confirmed. In order to determine their accuracy, this thesis georeferenced eleven historic maps and four plats to the modern landscape. As a result, possible locations of Charleston’s walled city were predicted. These were surveyed as test sites and remote sensing methods were employed that helped confirm or deny each hypothesis. After an analysis of georeferencing and ground penetrating radar results, the surveys of portions of the historic fortification system that was once made of earth and entrenchments were unable to identify with confidence any remaining features. However, the location of the brick-made curtain line on East Bay Street near South Adger’s Wharf and near South Market Street was identified, as well as a segment of the northern parapet wall of Craven’s Bastion located today under the United States Custom House.

Glacial Lake Agassiz is a paleolake that formed over central North America at theend of the last Ice Age at ~14.3 ka, and its drainage into the North Atlantic likelycontributed to the Younger Dryas cooling at ~12,900–11,700 cal yr BP. Drainage eventsare identified through the chronologic and geomorphic analyses of strandlines, includingbeach ridges, escarpments, and spits, which record past lake levels. Strandlines in a studyarea in the southern basin of Lake Agassiz in NW Minnesota warp upward toward thenorth as a result of glacioisostatic adjustment (GIA). Older strandlines have a steepernorth-south gradient than younger strandlines due to differential rates of rebound. Thestudy area contains a series of strandlines that are classified into 11 strandline groups(SGs) rather than named beach ridges previously identified along other sections of theLake Agassiz coastline. SGs are composed of one or two well-developed beach ridges, orsets of ridges, that branch out and increase in number toward the south, representingmany minor drops in lake level up to ~3 m due to a combination of incision of thesouthern outlet and uplift caused by GIA.A beach ridge formation model developed from littoral sediment sample analysesand GPR data at 100–500 MHz frequencies suggest that strandlines in the study areaformed in an overall depositional regressive system with intermittent storm events. Theestimated beach ridge rate of formation is ~8.5-17 yrs/ridge based on ~70 Lockhart Phasebeach ridges in the study area and the corresponding optically stimulated luminescence(OSL) dates for the Herman through Tintah beaches and meltwater signatures duringdrainage events. Strandline chronology is also assessed based on eight OSL datingsamples. Most of the dates fell within the error margin for published dates fromassociated strandlines at the southern outlet. However, samples from the Campbell andTintah beaches are approximately 3,000 years older than published dates. Two OSLsamples currently undergoing laboratory analyses may bring clarity to the differences inages.Strandlines were mapped in detail on a surficial geology map at a scale of1:24,000 as a USGS EDMAP project. A new lithologic unit (Qlsg) composed of silt,sand, and gravel that make up strandlines have aided in the analysis of strandlineformation in the southern basin of Lake Agassiz. Data analyses reveal that: 1) lake levelfall in the study area was overall continuous and gradual through time with many low-magnitudedrops in lake level, rather than few major episodic events, and 2) strandlinesbecome progressively younger with decreasing elevation lakeward.

This study analyzes the application and practicality of different non-destructive testing (NDT) methods for the inspection of reinforced concrete. There are two studies discussed in this paper. The first one is a laboratory investigation, utilizing ground penetrating radar (GPR) on concrete slabs that consisted of simulated defects to mimic those found in real concrete structures. This served as a preliminary experiment to verify the GPR data with the design of the slabs. The following study describes a field investigation on an old bridge deck, which involved multiple testing methods. The techniques used were GPR, impact echo (IE), ultrasonic pulse echo (UPE), a cover meter, and a rebound hammer. The primary method implemented was GPR for locating the embedded reinforcement steel and suspected areas of concrete deterioration caused by rebar corrosion. Core specimens were extracted for visual inspections to compare with the GPR data. Additionally, compressive tests were completed to obtain the compressive strength of the bridge deck. This data was compared to results from the rebound hammer, which provided an estimate on the in-situ compressive strength of concrete. The cover meter provided data on the rebar depth and spacing. Lastly, the IE and UPE instruments were utilized to determine the thickness of the bridge deck slab. Through data analyses, the results were compared primarily with the GPR and actual core data, in which most of the data was validated through an acceptable percentage accuracy and mean absolute error. This study displays the limitations when using NDT methods in a practical situation as opposed to laboratory settings. The uncertainty of what objects and flaws are embedded beneath an element poses some challenges; therefore, having the appropriate knowledge and background on NDT methods is necessary for structural evaluations. However, there is a promising future for NDT in civil engineering and construction project applications.

Nondestructive testing (NDT) is a reliable method for determining the structural integrity of new and old construction and understanding the condition of structural defects. It is widely acknowledged that the interpretation of nondestructive evaluation (NDE) has a significant impact on the reliability and consistency of this technology. However, NDT data acquisition remains subjective and heavily dependent on the operator’s knowledge and expertise. As a result, there are still issues to be resolved regarding the imaging and diagnostic procedures for NDT-based concrete inspection. NDT methods such as ground penetrating radar (GPR), and impact-echo (IE), have been extensively used to inspect concrete bridges for damage and deterioration. In this study, concrete slabs were investigated in a laboratory setting with simulated structural defects such as corrosion, delamination, honeycomb, and voids. To validate the accuracy and robustness of the two NDT methods, defects were embedded closely (ranging from 1 to 5 inches) and superimposed in two different layers for the same specimen. GPR B-scan (line scan) radargrams demonstrated hyperbolic shapes for voids and corroded rebars, whereas delamination was identified by a hollow strip pattern. GPR C-scan (area scan) data was used to calculate the size (length, width, and thickness) of delamination simulated by polystyrene sheets. The results showed an accuracy of 80% in predicting the thickness of delamination. GPR, however, did not detect delamination for objects smaller than 8 inches in size due to the dense meshing. The amplitude data from GPR A-scans (amplitude vs depth) were used to calculate the depth of a defect. Statistical measures exhibit a high degree of agreement between the depth measured by GPR and the actual depth of a defect. The IE method uses elastic waves to propagate through concrete for identifying the embedded defects. The interpretation of IE data involved post-processing of reflected signals in the time-domain and frequency-domain modes. The simulated structural defects did not show any discernible trend in time-domain data. However, defects were detected using IE thickness data from the peak frequency spectrum. The frequency-domain IE data was validated by comparing it to GPR data for similar concrete specimens.

The surficial aquifer system at the Wormsloe State Historic site has the potential to serve as an alternate groundwater source for non-potable water needs in Georgia. This study integrates ground penetrating radar (GPR), digital elevation model (DEM) derived from Light Detection and Ranging (LiDAR), hydrograph data, and porosity estimates to evaluate the total water storage and change in storage. The GPR data mapped the depth to the confining layer and the water table depth. The magnitude of seasonal fluctuations in hydraulic head was extracted from hydrograph data. These datasets were used to construct a water table elevation map for two seasonal conditions. Porosity estimate from the repacked samples was integrated and used to compute total water storage and the changes in storage. This study shows that the surficial aquifer is an alternate groundwater source but limited by seasonal fluctuation in water storage and lateral and inland saltwater intrusion.

Ground penetrating radar (GPR) is a valuable tool widely used for many tasks in infrastructure development and maintenance, mining and archaeology, as well as in support of construction industries. As opposed to the traditional monostatic GPR system, a multistatic subsurface synthetic aperture radar (SAR) system uses a fixed transmitter to illuminate the underground and a moving receiver to sample the scattered signals around the transmitter which provides major advantages in subsurface imaging and target detection. The most notable advantages are the faster acquisition time, the improved lateral resolution, and the higher gain resulting from coherent processing. Notwithstanding, the complexity of such a subsurface SAR system introduces new challenges which this dissertation aims to address. The first challenge is encountered in data generation using simulation. In order to evaluate the imaging performance of the subsurface SAR system, a new full-wave simulation is needed each time the receiver moves to a different position, which makes the number of simulations grow with the number of the required samples. By replacing the target with its equivalent current distribution and invoking the reciprocity theorem, this dissertation proposes a method that reduces the number of simulations required by a factor of N/2, where N is the number of samples required to form the SAR image. Another challenge is encountered in regard to the detection and localization of buried pipelines. Since pipelines have an extended geometry, the scattering phase center on a pipeline moves as a function of the transmitter and receiver positions. This renders traditional focusing algorithms, such as the back-projection algorithm, not applicable for such a target as the scattering phase center does not appear as a fixed point with respect to the positions of the radar's transmitter and receiver. Exploiting the theoretical knowledge for pipeline scattering behavior, this dissertation develops a dynamic focusing algorithm tailored to pipelines, thereby leading to their detection and localization in multistatic SAR mode. Building on the focusing algorithm for pipelines, which has limited range and cross-range resolutions, the dissertation provides a more general treatment for detection and localization of multiple closely-spaced pipelines. This is accomplished by developing a 3D forward scattering model for multiple pipelines, then inverting the model by minimizing the misfit between measured signals and the forward scattering model using a dynamic grid search algorithm. The third challenge addressed in the dissertation is related to the detection and localization of buried-pipeline leaks when using the multistatic SAR mode. Due to the multiple scattering-phase-centers phenomena exhibited by pipelines, the scattered signal from the pipelines tends to overwhelm the signal scattered from the leak, making leak detection a difficult task. Utilizing the pipeline detection and localization algorithm devised earlier, the dissertation suggests a technique to identify and eliminate the pipeline contribution of the radar response, leading to better performance in leak detection. Lastly, the problem of the direct signal leakage from the transmitter to the receiver in multistatic SAR is discussed. The leakage signal tends to be much stronger than the signals from buried targets. By incorporating array null-synthesis in the SAR processing stage, the dissertation proposes a method in which a null is created along the direction of the transmitter, thereby nulling out the leakage signal. This allows the signals from the buried objects to be detected, which ultimately leads to better imaging performance.

We develop the theory behind and a prototype of a 0.3-12 GHz ground-penetrating radar (GPR) system foruse in imaging the subsurface environment. Experimental results show that the prototype GPR system is likely able to perform in field applications including buried utility detection and humanitarian demining operations. First, we present two antenna designs – the resistively loaded vee dipole (RLVD) and elliptical RLVD based on a Guanella balun structure. However, we found a disconnect between the performance of these antennas in simulation and the real-world system performance. Therefore, we have developed a theory of how the less-often-considered aspects of antennas including coupling, distortion, beamwidth, and radar cross section dramatically affect the overall imaging performance. We perform a deep-dive into the causes, solutions, and design decisions that each aspect of the antenna requires in order to make it more likely that an antenna will perform well in the niche application of GPR. In order to validate and apply these principles in the real world, we present the creation of a 0.3-12 GHz GPR system in both the lab and outside environments. We show the entire build of the custom FPGA-based sampling receiver, examine the VHDL backend for data offloading, and describe the real-time visualization system. Next, we show in depth the signal-processing pipeline required to take raw data to a pre-processed image: de-glitching, de-jittering, and background removal algorithms, among others. The Kirchhoff-Migration imaging algorithms used to transform the pre-processed data into human-readable images are presented in detail. Finally, lab-based and real-world underground imaging results are presented as validation of the operation of the GPR system. We demonstrate the detection of steel rebar in a laboratory sandbox imaging test and portions of the underground structure around Kemper Hall at UC Davis.

Abstract: The population increase has stimulated the need for the creation and expansion of existing urban infrastructures such as sewer, water, power and telecommunication lines. In order to support this need, multiple efforts to find sustainable solutions that support the urbanization trend have been studied. As a solution, construction methods such as trenchless techniques—ranging from maxi horizontal directional drilling for large-scale pipe installations to microtrenching for the installation of telecommunications infrastructure, that offer reduced social and environmental impacts compared to conventional opencut construction methods, are becoming more accessible and increasingly used. Despite the advantages of underground construction, challenges exist in knowing the locations of existing subsurface utilities or other underground objects (such as rocks) an that uncertainty can impact new installations. Existing non-destructive technologies, such as ground penetrating radar (GPR), can be used to map extensive areas in a fast and accurate manner and support the construction of new infrastructure. However, GPR data is difficult to interpret, and requires an experienced person to be able to locate the features within the image that correspond to objects. In order to overcome the issue of data interpretation, multiple studies aiming towards the automation of GPR data interpretation have been conducted—however, the methods proposed can still improve in terms of detection speed and accuracy. The objective of this research is to automate GPR data interpretation to support underground construction. To achieve this, an extensive database of GPR images due to commonly encountered underground objects, such as rock-sized boulders and PVC and metal pipes, was collected based on GPR measurements in a laboratory setting. The database compiled from these measurements was used alongside different machine learning algorithms, including YOLO v3 and R-CNN, to determine the methodology that has higher accuracy of detection and classification for the automated object detection based on the correspondent features in GPR images. The limitations of the various algorithms are considered, and recommendations are proposed for future studies. A study on the long-term performance evaluation of micro-trenching backfill materials was continued, with GPR surveys performed to precisely determine conduit depth within the micro-trench. Updated measurements of conduit depth were compared to previously collected data to determine the conduit displacement four to seven years after installation.

Glacial Lake Maumee (an early ancestor to Lake Erie) developed extensive beach ridges and bars that included an elongate arcuate bar complex at Columbus Grove, OH, that was very similar to the modern Presque Isle flying spit at Erie, PA, in terms of shape, its 10 km length, its shoreline orientation, and its angle from the coast, although it is extremely narrow, distally discontinuous, and extremely thin. The bars consist of relatively fine-grained, poorly sorted, polymodal sediment that has been extensively bioturbated and plowed, although their shapes are well preserved on the modern surface and they retain useful grain size signals regarding depositional processes and settings. Lake floor sediments are primarily very fine to medium silts with secondary clays and minor sands. Bars appear to have accreted in place, locally with lateral expansion but without significant migration or erosion. Vertical successions vary from cryptic upward shoaling and coarsening to quite heterogenous (better seen by extracting peaks rather than using basic descriptive statistics). Bars are thin (commonly

Ground penetrating radars (GPRs) have been extensively used in many industrial applications, such as coal mining, structural health monitoring, subsurface utilities detection and localization, and autonomous driving. Most of the existing GPR systems are human-operated due to the need for experience in operation configurations based on the interpretation of collected GPR data. To achieve the best subsurface sensing performance, it is desired to design an autonomous GPR system that can operate adaptively under varying sensing conditions. In this research, first, a generic architecture for cognitive GPRs based on edge computing is studied. The operation of cognitive GPRs under this architecture is formulated as a sequential decision process. Then a cognitive GPR based on 2D B-Scan image analysis and deep Q-learning network (DQN) is investigated. A novel entropy-based reward function is designed for the DQN model by using the results of subsurface object detection (via the region of interest identification) and recognition (via classification). Furthermore, to acquire a global view of subsurface objects with complex shape configurations, 2D B-Scan image analysis is extended to 3D GPR data analysis termed “Scan Cloud.” A scan cloud-enabled cognitive GPR is studied based on an advanced deep reinforcement learning method called deep deterministic policy gradient (DDPG) with a new reward function derived from 3D GPR data. The proposed methods are evaluated using GPR modeling and simulation software called GprMax. Simulation results show that our proposed cognitive GPRs outperform other GPR systems in terms of detection accuracy, operating time, and object reconstruction.

MnDOT, as part of a multistate pooled fund project, is developing procedures that use ground penetrating radar for QA/QC in the placement of bituminous asphalt pavement. In July 2020, over 400,000 measurements were collected (using the GSSI PaveScan RDM 2.0) from a 5.5-mile stretch of US Highway 2 near Bena, Minnesota. In this thesis, we analyzed that data to develop conceptual models and identify statistical measures, methods, and tests to characterize the results. Graphical representations of our results are presented that will help guide the next phase of field testing with the ultimate goal of developing a national standard. Spatial correlation was characterized using the experimental semivariogram. This analysis revealed that the correlation length of the dielectric (and therefore air voids) ranges from 3[ft] to 45[ft] near the centerline joint, and is less than or equal to 13[ft] in the middle of the lane. These relatively small correlation lengths demonstrate a lack of significant spatial correlation in asphalt density measurements. Lateral variations in dielectric were revealed by partitioning measurements in the road into 1[ft] wide segments. Summary statistics were then computed using the measurements from each segment. The mean of the dielectric values from each segment varied throughout the road. This variation of the mean led to the development of a conjecture that splits the road into three unique zones: the joint zone (i.e., the centerline joint), the transition zone (i.e., the transition between the joint and the mat), and the mat zone (i.e., the middle of the road). This development is important, because current practices in transportation engineering only acknowledge the presence of two zones in the road (joint and mat). The possibility of sensor bias was examined through the use of histograms, probability plots, and statistical tests. Sensor bias was characterized by the prevalence of questionable measurements and dissimilar distributions for each sensor. In this case, the questionable measurements were identified as such because they were either not physically possible, or unlikely to represent an accurate road measurement. The dissimilar distributions demonstrate that, despite sampling the same population, each sensor makes significantly different measurements than the other sensors. Because it is likely present, sensor bias during data collection must be identified and addressed in real time. Recommendations for future data collection and analysis are also provided. These recommendations primarily pertain to sensor bias, data coverage, collection goals, and future experimentation. These recommendations allow MnDOT to move forward in the project knowing that testing procedures are statistically justified.

Common methods used to locate battlefield features have not been utilized extensively on the Charleston peninsula. This thesis explores methods used to locate battlefield features and their effectiveness on the Charleston peninsula. The methods were utilized in areas related to the Siege of Charleston in 1780 during the American Revolution. These methods include research into historic accounts of battlefield features, geographic information systems (GIS) to georeference historic maps of the Charleston peninsula, LiDAR to detect the topography of the land and locate any changes to the land over time, and finally ground penetrating radar (GPR) to locate any battlefield features. Through this study you will see that all these methods work together well to locate battlefield features. The final method, GPR, located the remains of the earthwork fortification in Wragg Square, the tabby hornwork in Marion Square, and the foundation of the Cistern Yard barracks on the College of Charleston campus. These combined methods located areas related to the Siege of Charleston that had not been mapped.

Raised beach ridges in polar environments are recorders of relative sea-level (RSL) change. The South Shetland Islands (SSI) contain large areas of ice-free raised beaches used to resolve paleo-sea-level change. The largest of these ice-free areas is Byers Peninsula (60 km2) on Livingston Island, SSI. This peninsula and other ice-free coasts within the SSI contain series of raised beach ridges correlated by age and elevation. The internal architecture of these beach ridges is critical to understanding paleo-RSL. The non-destructive method of ground penetrating radar (GPR) was used to investigate the internal architecture of the South Beaches of Livingston Island where a flight of nine beach ridges (five of which were the focus of in this study) have been glacio-isostatically raised since the last glacial maximum (LGM). Recalibrated radiocarbon ages collected from these and other beach ridges in previous studies are used to relate beach morphology and stratigraphy to the glacial history of the region. Over 10 km of 200 MHz GPR transects and synchronous GPS were collected over a two-week period in the spring of 2019. Sediment pits

Concrete structures experience different kinds of mechanical loadings, physical and chemical interactions, and aggressive environmental attacks during their lifetime. It is crucial to perform regular monitoring of these structures to ensure their functionality and to provide public safety. Furthermore, the selection of the most reliable testing methods for the overall condition assessment is also important. Thus, this research is focused on finding the structural condition of a precast deck section in the Lincoln Parking Deck at YSU using a Ground Penetrating Radar (GPR). The main objective of using a GPR is to identify its capability for the condition assessment of concrete structures. A GSSI SIR 4000 mainframe system with a 1.6 GHz antenna was deployed for the radar survey. Small concrete slabs were prepared in the laboratory with fresh and corroded reinforcing bars. Corrosion in reinforcing bars was artificially developed using a Q-Fog cyclic corrosion tester. The GPR response from the slab, having the fresh rebar at varying depth, indicated that the reflection amplitude obtained from the scan of the rebar target decreases with increasing depth. Likewise, comparing GPR B-scan from the fresh and the corroded rebar indicated a qualitative and a quantitative difference in the result. The qualitative difference was observed as the hyperbolas (rebar reflections) showed weakening smoothness, brightness, and visibility with the increase in the amount of corrosion. In the same way, the quantitative difference was observed as there was a decreasing trend in the reflection amplitudes of each point of a hyperbola with an increasing amount of corrosion. The conclusion made from a laboratory experiment was applied for developing the corrosiveness map of a precast deck.The field experiment was performed using a three-wheeled antenna cart setup and following the standard bridge deck survey procedures. Information on rebar size, spacing, and depth was extracted by quantifying the GPR B-scan response. Both numerical and image-based analysis methods were followed for developing the corrosiveness maps of the precast deck section. The corrosiveness map showed a few critical sections in the precast deck. For further analysis, a visual inspection of the precast deck was carried out. A few cracks, delamination and rust scales were visible around the expansion and construction joints of the precast deck. The location of these defects in the precast deck matched the location of the critical sections in the corrosiveness map. The present condition of a precast deck section was found satisfactory except for a few areas of serious cracking that need repairs to control the further growth of corrosion. Furthermore, GPR results showed higher accuracy with the field inspection outcomes demonstrating that a GPR can be used as a reliable non-destructive tool for the condition assessment of concrete structures.

The goal of this research was to develop an algorithm to process measured data using an ``optimal'' geometry in order to reconstruct an image of weak targets when strong scatterers are present. This thesis is focused on Weak Target Detection and Imaging in Radio Frequency Tomography (RFT) for Ground Penetrating Radar (GPR). We first start by simplifying the solution of the inversion problem in Radio Frequency Tomography (RFT) based Ground Penetrating Radar (GPR) in order to minimize the compute time for image reconstruction of shallow buried objects. A further goal is to increase the resolution and sharpen the quality of the reconstructed image as we concentrate on weak target detection and imaging in RFT. We propose an accurate, fast method to reconstruct the image of below ground targets using an optimal linear filter, such as matched filter processing. Moreover, a novel method is proposed to improve RFT to detect and imaging weak targets surrounded by strong scatterers in measurement domain. The match filter is the most common approach used to simplify the solution of the inversion problem in GPR model. The proposed method increases the signal-to-noise ratio (SNR) to enhance the quality of the image. Using this technique leads to a decrease in the reconstruction time. Also, it reduces the data acquisition time which is critical in many commercial GPR applications. Compared with other inversion algorithms such as truncated singular value decomposition (TSVD), matched filter algorithms yield a high quality 2D image of shallow buried objects with minimal computational load or noise effect.However, in case of Multi-Target Detection (MTD), RFT has an inherent weakness, especially when strong interfering scatterers and weak targets are present in the same measurement domain. Strong sidelobes from dominant (strong) scatterers can interfere with the echoes from weak targets, thereby leading to missed target detections, and a decrease in the quality of the reconstructed image. Moreover, these sidelobes form electromagnetic fields in the measurement domain with characteristic patterns centered on strong scatterer. These electromagnetic fields decrease the quality of the reconstructed image and can mask weak targets close to the strong targets in the region of interest. In this work, we proposed a faster, more flexible and more accurate technique to decrease the effect of strong sidelobes on weak targets by modeling the strong scatterers as ``pseudo'' transmitters, using the information from the Dyadic Contrast Function (DCF) to represent the scene under investigation. DCF is analyzed in order to remove the effect of these strong sidelobes.This approach uses a ``Suppression Algorithm (SA)'' to remove the effect of strong sidelobes on weak targets, and to reconstruct images using DCF analysis. The proposed algorithm starts by detecting the dominant scatterers in the reconstructed image domain. Then, each dominant cell in the reconstructed image represented as a DCF using its eigenvalues and eigenvectors to model the strong cell as a pseudo or secondary source. The electromagnetic characteristics, magnitude and phase, of the new source (transmitters) are estimated from the eigenvalues and eigenvectors for each cell in the investigation domain. The magnitude of the largest eigenvalue from the DCF will estimate the magnitude of the pseudo transmitter, while the eigenvector provides information about the orientation of the source. Preforming the eigen-decomposition of the DCF improves the process of locating, imaging, and identifying shallow weak objects. Furthermore, to simplify the inversion problem in the complete forward model, Iterative Reconstruction Algorithms (IRT), such as, Subsurface Multiplicative Algebraic Reconstruction Techniques (SMART) are considered as Row Action algorithms (RA-algorithms), providing faster, more flexible and better accuracy for solving the set of linear systems with non-square matrices and ill- conditioned linear operators. The SMART suppression algorithm offers an approach by modeling the strong scatterers as ``extra'' or pseudo dipoles (transmitters) in the measurement domain. The sidelobes of these dipoles are subtracted from the image to sharpen the quality of the weak target return. The presented algorithm has been verified using simulated RFT data, generated by the computational electromagnetic software FEKO, for regular and irregular targets scenarios. Our research shows, that using information from the DCF, it is possible to obtain high quality imagery of buried weak targets in RFT.

This thesis proposes a fast matched-filtered based imaging algorithm to detect and image below ground objects with ground penetrating radar. To image below ground objects, a set of distributed transmitters and receivers are placed above the ground, or slightly buried. The transmitters radiate waveforms into the subsurface. The resulting wave-front impinges upon underground objects, scattering electromagnetic energy in all directions. Receivers collect the reflected electromagnetic signals, retrieve the phase of the scattered signals, and transmit this information to a signal processing system for post-processing. After applying adaptive signal processing algorithms to the collected data, an image of the buried objects can be reconstructed. Reconstructed 2D images of buried objects are computed via numerical discretization and matched filtering techniques. The matched filtering technique requires less computational power and is easier to implement as compared to the matrix inversion technique. Results from simulation analyses and experimental data are used to validate this method. This work includes simulations and experiments with ground penetrating radar calibration techniques to justify the method introduced here.

This work focuses on the operational and safety issues associated with karst voids in large opening underground mines. Issues include water inrush, structural instability, and engineering uncertainty in these environments. Coupled with the fracturing prevalent in folded sedimentary rocks, karsts are complex and challenging ground control risks. Traditional methods of predicting karst void locations such as probe-drilling are impeded by the inconsistent spatial distribution and variable sizes of the features. Ground penetrating radar (GPR) is a geophysical technique that transmits radio waves into a medium and subsequently detects reflected waves via a receiver. The travel time and energy of received signals are then processed and interpreted. The difference in material properties between limestone and open karst voids causes strong reflections. This work summarizes a series of 2D and 3D GPR surveys for karst void mapping within a mine pillar and within sill pillars between mine levels in a large opening underground limestone mine. In this case study mine, karst voids are hazardous ground control risks that interact with geologic discontinuities, creating free blocks within the rock mass. As tunnels are advanced via blasting, unknown karst voids may be exposed and pose risks to mining personnel. The karst voids also form a hydrogeological network of water reservoirs with spatial locations throughout the rock mass that are difficult to predict with traditional methods such as drilling. While GPR has been utilized throughout several industries for anomaly detection, mapping, and validating other geophysical data sets, this technique has not seen the same proliferation within the mining industry. Regarding published literature, there is a lack of works that detail the applicability of GPR in underground mining scenarios. The aim of this work is to expand on previous methodologies establishing GPR as a useful tool in underground mining applications, and to discuss the benefits and limitations GPR data in such scenarios.

Sea level rise as a function of climate change is projected to have profound impacts on coastal environments, and groundwater inundation is expected to follow. In this study, Bryan County, Georgia (USA) was evaluated using ground penetrating radar to determine future impacts of groundwater inundation to onsite wastewater treatment systems (OWTS). Also known as septic fields, OWTSs were mapped using handheld GPS units. These locations were then compared to groundwater projections of 0.2, 0.5, and 1.0-m heights above the seasonal high-water table. OWTS depths were modeled at 0.45 and 0.9 m, based on similar studies in North Carolina (Manda et al., 2015). These conditions create an environment where OWTSs will be a danger to water resources in the area. Modeling shows OWTS groundwater inundation ranging from 99% in Upper Bryan County to 100% in Lower Bryan County at seasonal high-water tables and 1 m of groundwater rise.

To augment training data for machine learning models in Ground Penetrating Radar (GPR) data classification and identification, this thesis focuses on the generation of realistic GPR data using Generative Adversarial Networks. An innovative GAN ar- chitecture is proposed for generating GPR B-scans, which is, to the author’s knowledge, the first successful application of GAN to GPR B-scans. As one of the major contri- butions, a novel loss function is formulated by merging frequency domain with time domain features. To test the efficacy of generated B-scans, a real time object classifier is proposed to measure the performance gain derived from augmented B-Scan images. The numerical experiment illustrated that, based on the augmented training data, the proposed GAN architecture demonstrated a significant increase (from 82% to 98%) in the accuracy of the object classifier.

Unsaturated flow in the vadose zone often manifests as preferential flow resulting in transport of water and solutes through the soil much faster than would occur for uniform matrix flow. Time-lapse ground penetrating radar (TLGPR) shows promise as a non-invasive means to monitor unsaturated flow and here is used to monitor lab-scale forced infiltration events for capturing evidence of non-uniform and preferential flow phenomena directly from arrivals in the GPR images while simultaneously characterizing parameters of the flow system, such as bulk water content and rates of wetting front movement. This was accomplished by 1) directly interpreting transient arrivals in GPR profiles for evidence of ono-uniform flow and 2) with the aid of migration processing techniques to improve the quality of GPR images for identification and tracking of transient arrivals related to wetting in the soil. A novel method is described and evaluated to characterize the 2D velocity structure of a soil and used to migrate the GPR images. This method incorporates multi-offset measurements to characterize the depth to a potentially unknown static reflector and root mean square (RMS) velocity above the reflector with incremental changes in travel time to the static reflector and a transient reflector (i.e. the wetting front) determined from single-offset constant offset profiles to determine incremental changes in velocity above and below the transient arrival. The method is applied to TLGPR data during infiltration experiments in a 60 cm deep sand-filled tank and monitored with water content probes. To verify the approach the methodology is applied to GPR data simulated using transient water contents generated by the unsaturated flow simulator HYDRUS 2D given lab-measured hydraulic properties of the soil. For both the empirical and simulated data, we found that the 2D velocity analysis was effective in monitoring changes in the wetting front and that migration of the reflection profiles was able to improve the interpretation of non-uniform flow.

Imaging of visually obscured objects within buildings or buried under ground using low frequency EM waves have many military, security, civilian, and industrial applications. In imaging systems, range and cross-range resolution are important parameters that determine the ability of the system to resolve closely spaced objects. Range resolution generally depends on the signal bandwidth while the cross-range resolution depends on the antenna aperture size. In certain application such as through-the-wall imaging and detection of buried land mines and pipelines where the objects are closely spaced and/or located at far distance from the transmitter and receiver, the imaging system must have high cross-range resolution. Typically, this is achieved by using a large array of directive antennas with limited field of view and low mobility. In this research, new imaging methods based on synthetic aperture processing are presented to address low cross-range resolution, limited field of view, and low mobility of currently available imaging systems. For through-the-wall imaging applications, the large array of directive antennas used in current systems is replaced by a small number of omnidirectional transceivers mounted on moving robots. One or more receivers move and sample the scattered signal at different locations within the imaging area. By applying back-projection technique to the obtained samples, a large 2-D array is synthesized for high resolution imaging. Using omnidirectional antennas, images with a 360° field of view are generated in a short period of time. To suppress the effect of direct coupling between the transmitter and receiver and multiple reflections in the image, orthogonal circular polarizations are used for the transmitting and receiving antennas. In this research, a novel wideband circularly polarized omnidirectional antenna is presented for through-the-wall imaging applications. The antenna operates based on excitation of orthogonal field distributions similar to a circular waveguide TE21 mode. A bi-static FMCW radar system realizing the presented through-the-wall imaging method is designed and fabricated. The system is configured to account and completely compensate for the delays inherent to the circuits and errors in transmitter/receiver synchronization completely. Performance of the method is evaluated in different real-world scenarios using the fabricated radar system and the associated algorithms. Two new methods for image enhancement in through-the-wall imaging are developed. The first method enhances the range resolution and reduces the background noise in the image by detecting the locations of the reflections and forming the image only at those points. The second method discriminates wall surfaces from small size objects and detects the location and orientation of all wall surfaces within the imaging area. The second part of the dissertation describes a new method for subsurface imaging using the general idea developed for the through-the-wall imaging. The method utilizes bi-static ground looking transmitters and receivers mounted on moving robots to sample scattered fields at different locations. The samples are processed coherently to form a large 2-D synthetic array which provides high cross-range resolution 3-D images of the subsurface in a short period of time. This cannot be achieved using conventional ground penetrating radars. A new low profile wideband antenna is designed for the imaging of deeply buried targets or targets buried in soil with high losses. Performance of the designed antenna and the proposed imaging method have been tested through field measurements. The imaging results show high resolution 3-D imaging capability of the method.

In this dissertation, we explore methods to combine multiple frequency Ground Penetrating Radar (GPR) signals in a manner to improve the resolution of images of deeply buried targets. We propose using an optimization problem solver to combine multiple GPR frequency scans over the same area to improve image resolution. First, we discuss GPR basics. Second, we report on a method to simulate GPR radar scans over any type of terrain, any frequency, and any target depth for use in our study of GPR compositing methods. Third, we define an optimization problem solver, exploring its capability to achieve reasonable results as well as propose a figure of merit for the best solution measurement tool. In comparing the optimization problem solver result to methods previously explored in the literature, detailed by Dougherty [8], then Booth [5][48] and finally Bancroft [3], we found our algorithm exhibits a meaningful improvement compared to the named methods. As an extension, we explored comparing scans from various heights using the optimization problem solver method with a Chirp excitation function at the same heights, finding edge detection improved with the response from the Chirp excitation function, but depth detection poorer than the optimization problem solver.

This thesis focused on classifying GPR cylinders' B-scans according to their depth, size, material, and the dielectric constant of the underlying medium using four different architectures of convolutional neural networks. Two CNNs were newly proposed for this study, while the other two were used by other authors. These CNNs were trained using a couple of adjusted training options including initial learning rate, learn rate drop factor, and learn rate drop period; which had a positive impact on a part of the used models, while the option maximum number of epochs worked good with all of the used models. Results show that the first newly proposed CNN showed a superior performance due to the use of a deep network with a large amount of small filters. Using this model, it was found that the best results were carried out when GPR B-scans were classified according to the cylinders' materials.

This research investigated alternative strategies for making buried non-metallic pipes (CFRP, GFRP, and PVC) easily locatable using Ground Penetrating Radar (GPR). Pipe diameters up to 12" and buried with up to 4 ft. of soil cover were investigated. The findings of this study will help address the detection problem of non-metallic pipelines and speed the adoption of composite pipes by the petroleum and natural gas industry. The research also investigated the possibility of locating buried pipes transporting hot fluids using Infrared Thermography (IRT). Results from the study have shown that, using carbon fabric and aluminum tape overlay on non‑metallic pipes (GFRP or PVC for this study) before burying significantly increases the reflected GPR signal amplitude, thereby making it easier to locate such pipelines using GPR. The reflected GPR signal amplitude for pipe sections with carbon fabric or aluminum foil overlays was found to have increased by a factor of up to 4.52 times, and 2.02 times on average across all the pipe sections tested, from the baseline (unwrapped) pipe sections. The research also highlights the importance of using the correct antenna frequency for detecting buried pipes in wet soil conditions. Wet soils with high electrical conductivity and dielectric constants have higher radar signal attenuations that significantly affect the penetration depth and returned signal amplitudes from buried objects. A 200 MHz frequency antenna was found in this study to be ideal for locating the buried pipes in all soil moisture conditions. The 200 MHz antenna was able to detect buried pipes up to the maximum 4 ft. depth of soil cover that was studied experimentally. Numerical estimation using the same soil from the experiment shows that this antenna can penetrate up to a depth of at least 5.5 ft. in very wet clay soils with volumetric water content of 0.473. After evaluating the attenuation characteristics of different radar antennae, it was found that material/ohmic attenuation is constant across a range of antenna frequencies; the increase in GPR signal attenuation associated with higher antenna frequencies was found to be a result of scattering attenuation from subsurface inhomogeneity/clutter. Scattering attenuation is however usually ignored in literature, resulting in erroneous estimation of radar signal attenuation. Finally, laboratory study proved that, heat from a buried pipeline transporting hot fluid can propagate through the soil to the surface and be detected using IRT. Additionally, a 6" diameter steam pipe with a 6" minimum insulation and buried with 2.5 – 3 ft. of soil cover was easily detected in varying soil moisture conditions during different seasons throughout the year using IRT in the field environment. The successful application of IRT in detecting this pipe proves the potential for using this technique in locating buried pipes transporting hot fluids such as steam or petroleum products from production wells or refinery plants.

Paleochannels preserved on terraces via meander cutoffs during an incisional period record the channel geometry and thus discharge at distinct points in time throughout a river’s history. We measured paleochannel geometry on terraces throughout the Le Sueur River in south-central Minnesota, to track how channel geometry has changed over the last 13,400 years. A rapid drop in base level 13,400 yr B.P. triggered knickpoint migration and valley incision that is ongoing today. Since the 1800’s, the area has developed rapidly with an increase in agriculture and associated drainage, directly impacting river discharge by increasing water input to the river. Five paleochannels were identified on terraces along the Le Sueur River from 1m-resolution lidar data. Ground Penetrating Radar (GPR) was used to obtain a subsurface image across paleomeanders to estimate the geometry of paleochannels. Paleochannel geometry and estimated discharge were then compared to modern conditions to assess how much change has occurred. Three lines were run across each paleochannel perpendicular to the historic water flow. Each of the 15 lines were processed using the EKKO Project 2 software supplied by Sensors and Software to sharpen the images, making it easier to identify the paleochannel geometry. Paleodischarge was determined using the Law of the Wall and Manning's Equation, using modern slope and roughness conditions. OSL samples were collected from overbank deposits on terraces to determine the time of channel abandonment, and supplemented with terrace ages obtained from a numerical model of valley incision. Paleodischarge coupled with depositional ages provide a history of flow conditions on the Le Sueur River. Results show an increase in channel widths from the time paleochannels were occupied to modern channel dimensions from an average of 20 meters to 35 meters. The change was not constant through time, as all paleochannels analyzed on terraces had similar-sized channels. The best way to determine paleogeometry was using the 'best interpretation' of GPR data couple with coring data; and paleodischarge was best estimated using Manning's equation with an n value of 0.035. Results show an increase in discharge compared to paleochannels of a factor of two. Uncertainty estimates in GPR-based paleogeometry can change paleodischarge calculations by 50 %. Incremental flood frequency analyses, based on data obtained from the Red Jacket stream gage at the outlet of the Le Sueur, suggest a 1.5- and 2-year flood of 102 m3/s and 154 m3/s, respectively, which is comparable to estimations of bankfull based on current channel geometry at the Red Jacket gage, validating the methodology. Problems associated with paleogeometry estimations are primarily due to meander bend preservation in the subsurface, challenging GPR interpretation. The increase in channel geometry and discharge implies that the increase in flow associated with drainage and climate change since the area’s development has greatly impacted the Le Sueur River. This resulted in a change in channel morphometry through increased erosion along the bluffs and banks, widening channels. This increase in erosion has directly impacted the amount of sediment delivered to the rivers from banks and bluffs, increasing the fine sediment load in this turbidity-impaired river system.

Radio Frequency (RF) Tomography is a mathematical process of 3D image reconstruction from a measurement using a multistatic distribution of transmitters and receivers. The geometric diversity of these elements increases the information in the measurements. The process of determining the permittivity and conductivity profile in the measurement domain, and, therefore, the shape of the target, from the scattered field measurements, is an inverse problem. To solve this problem, under conventional methods such as the Born approximation, we use the principles of linear scattering to determine a linear relationship between measured returns and target shape. The Born approximation is valid if the scatterer is small and does not interact strongly with other objects. However, strong scatterers within the domain may generate sidelobes masking weaker returns. This masking, in conjunction with multipath effects, may result in loss of features and subsequent failure to identify a target. In this research, a novel method is proposed to increase overall image quality and extend the capabilities of RF tomography by modeling the strong scatterers in the measurement domain as dipoles that behave as secondary sources (transmitters). Unlike conventional methods, the dipole model reduces the effects of the sidelobes from the strong scatterers and exploits the multipath of multiple targets or complex shapes. The multipath phenomena contains more information about the targets permitting illumination in the shadowed region and an increase to the radar aperture length. The electromagnetic characteristics for each modeled dipole are estimated by representing the cells in the measurement domain's image. The eigenvalue and eigenvector from each cell represent the phase and magnitude for the modeled dipole and also the spatial orientation of the target. The process of modeling large scatterers as dipoles can be iterated, addressing one strong scatterer at a time. This method effectively suppresses the sidelobes and exploits the multipath within the measurement domain. Using the Born approximation, the linear relationship between the scattered fields and the target is updated for simplicity. With iterations, the “extra” dipole will account for the multipath effects, thus removing some limitations caused by the Born approximation. This concept has been successfully demonstrated in software (FEKO© by Altair). In addition, this work also presents an innovative conversion using a back-projection algorithm for multipath effects and modeling of an “additional” source or transmitter in the measurement domain. The result of implementing this method of modeling strong scatterers as dipoles successfully demonstrated an increase in the resolution and enhanced radar imagery.

Ground Penetrating Radar (GPR) research conducted in coastal environments is one area that is lacking in archaeology. Surveys conducted in this type of environment afford the opportunity to evaluate the practical use GPR under field conditions. Coastal environments are effective for this evaluation because they offer a host of conditions that GPR surveys do not normally encounter at one time. The relationship of the land to the coast, sub-surface conditions and reliable survey areas create a "perfect storm" to test how practical the use of GPR is in coastal environments. This research is a study of homestead cemeteries situated within the boundaries of Cape Canaveral Air Force Station (CCAFS), using GPR. The research has three main goals. The first is to utilize GPR to identify if there are any unknown burials at CCAFS. The second is to test the practical effectiveness of GPR in coastal environments where high water table, geology and saline conditions can limit the capability of the technique to resolve subsurface features. The third is to correlate data from the GPR survey with ethnographic information to enhance the protection and maintenance with what is already available for the cemeteries. Research methods include field-based geophysical data collection in addition to archival and ethnographic historic research. The field component, to which this research pertains, entailed an on-site GPR survey at the nine sites on Cape Canaveral Air Force Station. This was followed by analysis of the information from the survey using standard processing software. Subsequently, a thorough archival search was completed to link historic and ethnographic information with the archaeological data obtained on the cemeteries. The final result of this research was a report that provides a detailed description of the results of the GPR survey of the cemeteries at Cape Canaveral Air Force Station.

Ground penetrating radar (GPR), electrical resistivity and seismic refraction surveys wereused to image possible buried sinkholes and identify potential areas of subsidence in the shallowpaleokarst surface of the Columbus Limestone at Ohio Caverns in Champaign County, Ohio. Aburied sinkhole, incipient sinkholes and a possible buried cave passage spatially correlated withorthoimagery and surface evidence of subsidence. Correlations were established by using GISsoftware overlaying a structure-contour map of the paleokarst, as determined by GPR data, overorthoimagery of the area. Wenner and dipole-dipole electrical resistivity surveys suggest regionsof lower resistivity surrounded by higher resistivity are associated with suspected subsidence.Refraction seismic data, collected along the same transect as the resistivity surveys, alsocorrelated with an area of depression in the region of a suspected sinkhole. The seismicvelocities from the refraction survey indicated lower depths for the clay soil and Ohio Shalecontact and also the boundary between the Ohio Shale and the Columbus Limestone. Remotesensing using electrical resistivity, GPR and seismic refraction techniques successfully imagedthe stratigraphy of the area and suggest other areas of incipient sinkhole formation.

The root systems of forest ecosystem are both important and understudied. Roots function as the critical organs of nutrient uptake and anchor tree stems against wind throw and aboveground disturbance. Roots are also an important terrestrial reservoir of carbon and play an important role in global climate. Conversion or disturbance of forest lands can result in an emission of carbon dioxide to the atmosphere, and so quantifying the amount of carbon in root biomass is of interest for forest scientists and the climate policy their research informs.This dissertation contributes to the science of root biomass measurement and modeling, providing techniques and approaches to improve those estimates, and models for leaf area that can be used to estimate root biomass. It focuses on techniques to improve measurement of root biomass using ground penetrating radar, an emergent method to measure root biomass. First, I investigate what above-ground metrics might best be used to model and predict root biomass and distribution and I develop new models relating basal area to leaf area for quaking aspen (Populus tremuloides) in the Sierra Nevada. Second, I investigate root biomass and distribution in coast redwood (Sequoia sempervirens), and how water acquisition may drive the allocation to roots in redwoods.Ground penetrating radar is shown to have considerable utility for the measurement of coarse root biomass. Above ground biomass is shown to be an excellent predictor of coarse root biomass in Sierran ecosystems. Leaf area, which is predicted by pipe model theory to have a tight relationship with fine root biomass, is shown to have some predictive power for root biomass, and my model represents an improvement over other reported metrics. Finally, I conclude that global root biomass estimates have not been based on reliable or unbiased data to date and I contribute A plan to correct this gap in future global estimates.

Researchers have studied the subglacial and proglacial impacts of glacial outburst floods through observation and measurements during flooding events and through analysis of the landscape and sediments after flooding events. While both erosion and sediment deposition dynamics have been studied extensively (eg. Sharp, 1985; Rijsdijk et al., 1999; Van Der Meer, 1999; Overgaard and Jakobsen, 2001; Glasser et al., 2003; Russell et al., 2006;); the routing and effects of pressurized groundwater during glacial outburst flooding are currently poorly understood.The objectives of this research are; 1) to test whether or not Ground Penetrating Radar (GPR) can can effectively image nearly vertical subsurface sedimentary structures thought to result from hydrofracturing of consolidated and confined glacial outwash sediments; 2) to determine if these subsurface sedimentary structures are in fact derived from pressurized groundwater.The site of this study is on western Skeidararsandur, Iceland, near the Sula river. Skeidararsandur is the largest glacial outwash plain on Earth encompassing an area of approximately 1300 km2 (Klimek, 1973). Because it is subject to frequent glacial outburst floods (Roberts et al., 2002) it makes an excellent laboratory to study glacial outwash structures. There are many surface ridges that have been interpreted as groundwater escape structures and correlated to previous outburst floods (Russell et al., 2006; Munro-Stasiuk et al., 2009). In addition, clastic dykes appear to represent groundwater escape pathways through the sands and gravels of the sandur (Munro-Stasiuk et al., 2009). GPR was able to detect steeply dipping clastic dykes and direction of flow at the study site. In addition, GPR can detect sinuous subsurface linear geologic features using proper scan orientation. The results of this study will lead to a better understanding of the dynamics of pressurized groundwater escape mechanisms resulting from sudden glacial outburst flooding. They will also help determine how pressurized groundwater moves through, interacts with, and responds to the sedimentary environment. Understanding how pressurized groundwater behaves will allow us to design better abutments, foundations, cores, and liners capable of withstanding immense stress. GPR gives us a non-destructive and inexpensive method that can be used to a depth of several meters for surface scans and as deep as needed in bore hole scans.

Accurate characterization of subsurface fractures is indispensible for contaminant transport and fresh water resource modeling because discharge is cubically related to the fracture aperture; thus, minor errors in aperture estimates may yield major errors in a modeled hydrologic response. Ground penetrating radar (GPR) has been successfully used to noninvasively estimate fracture aperture for sub-horizontal fractures at outcrop scale, but limits on vertical and horizontal resolution are a concern. Theoretical formulations and field tests have demonstrated increased GPR amplitude response with the addition of a saline tracer in a sub-millimeter fracture; however, robust verification of existing theoretical equations without an accurate measure of aperture variation across a fracture surface is difficult. The work presented here is directed at better verification of theoretical predictions of GPR amplitude and phase response. For sub-vertical resolution features, the response of a 1000 MHz PulseEKKO Pro transducer to a fluid-filled bedrock fracture analog composed of two plastic (UHMW-PE) blocks was measured, where fracture aperture ranged from 0-40 ± 0.3 mm and fluid conductivity from 0-5700 ± 5 mS/m. The GPR profiles were acquired down the centerline of the block, horizontally stacked to reduce errors, normalized to the control response at zero aperture, used to calculate reflection coefficient by dividing by the magnitude of the direct wave, and used to calculate the instantaneous phase. For sub-horizontal resolution features, lateral fracture extent ranged from 0-20 cm and fluid conductivity from 20-5700 ± 5 mS/m. GPR profiles were acquired parallel and perpendicular to the fracture. Comparison of the measured GPR response to analytical and numerical modeling suggests that numerical modeling best predicts both amplitude and phase variations due to changes in fracture aperture and conductivity. The Widess equation combined with an empirically derived scaling factor also predicts GPR amplitude response but not phase. Future applications to inversions of field data to map subsurface fracture networks will rely on easily invertible models, and numerical modeling using GPRMax2D can help develop a theoretical model for computationally effective and accurate inversion.

Currently there is no systematic quantitative methodology in place for the integration of two or more coincident data sets collected using near-surface geophysical techniques. As the need for this type of methodology increases—particularly in the fields of archaeological prospecting, UXO detection, landmine detection, environmental site characterization/remediation monitoring, and forensics—a detailed and refined approach is necessary. The objective of this dissertation is to investigate quantitative techniques for integrating multi-tool near-surface geophysical data to improve subsurface imaging and reduce uncertainty in discrete anomaly detection. This objective is fulfilled by: (1) correlating multi-tool geophysical data with existing well-characterized “targets”; (2) developing methods for quantitatively merging different geophysical data sets; (3) implementing statistical tools within Statistical Analysis System (SAS) to evaluate the multiple integration methodologies; and (4) testing these new methods at several well-characterized sites with varied targets (i.e., case studies). Three geophysical techniques utilized in this research are: ground penetrating radar (GPR), electromagnetic (ground conductivity) methods (EM), and magnetic gradiometry. Computer simulations are developed to generate synthetic data with expected parameters such as heterogeneity of the subsurface, type of target, and spatial sampling. The synthetic data sets are integrated using the same methodologies employed on the case-study sites to (a) further develop the necessary quantitative assessment scheme, and (b) determine if these merged data sets do in fact yield improved results. A controlled setting within The University of Tennessee Geophysical Research Station permits the data (and associated anomalous bodies) to be spatially correlated with the locations of known targets. Error analysis is then conducted to guide any modifications to the data integration methodologies before transitioning to study sites of unknown subsurface features. Statistical analysis utilizing SAS is conducted to quantitatively evaluate the effectiveness of the data integration methodologies and determine if there are significant improvements in subsurface imaging, thus resulting in a reduction in the uncertainty of discrete anomaly detection.

It is well known that a finite-size scatterer has a set of natural resonances, which are uniquely determined by the physical properties of the scatterer. This is also the case for a breast tumor which can be regarded as a dielectric scatterer. Since the scatterer is naturally "tuned" at the resonances, it is expected that an increased electromagnetic coupling would take place at the resonance frequencies compared to other frequencies. For a breast tumor, this would mean a higher power absorption, indicating a faster temperature increase resulting in more efficient hyperthermia. In this dissertation, an adaptive microwave concept is demonstrated for breast cancer applications. The general approach is to detect and identify the tumor-specific resonance, determine the electrical location of the tumor, and apply the focused microwave hyperthermia using the identified resonance and the electrical location. The natural resonances vary depending on the tumor size, shape, and breast tissue configuration. Therefore, an adaptive tuning of the microwave source to tumor-specific resonance frequencies could improve the overall efficiency of hyperthermia treatment by allowing for a faster and more effective heating to achieve a desired therapeutic temperature level. Applying the singularity expansion method (SEM), both the resonances and the electrical location can be obtained from the poles and residues, respectively. This SEM-based approach is computationally inexpensive and can easily be implemented as a combination processing into emerging UWB microwave systems. Alternatively, a relatively simple microwave system based on this concept can potentially be used in conjunction with existing mammography.

Abstract: Ultra-wideband (UWB) pulses scattered from objects contain essential information. By analyzing the scattered pulse, the objects can be located, imaged and characterized. This thesis focuses on characterizing physical and electrical properties of planar metallic and dielectric objects by analyzing the scattered UWB pulse. In the analysis of UWB pulse scattered from planar metallic objects, we establish an analytical model to estimate the scattered UWB pulse. Furthermore, we formulate a frequency-averaged radar cross section (RCS) to estimate the radar signature of an object based on UWB radar measurements. In the analysis of UWB pulse scattered from planar dielectric objects buried in ground, we establish a numerical model, based on mean square error (MSE) method, to estimate the depth, thickness and complex permittivity of the buried objects. Multiple experiments validate the proposed methodologies of analyzing the scattered UWB pulse from the planar objects, providing agreement between the measured results and proposed theory.

The purpose of this study is to extend the use of ground penetrating radar methodology towards a more reliable and accurate interpretation of pavement conditions. First, a complete set of 3D layered electromagnetic Green's functions is derived by way of transverse electric and transverse magnetic scalar potentials, featuring a new "direct" formulation for the field forms of the spectral Green's functions. The improper integrals underpinning the computation of the corresponding point-load solutions in the spatial domain are evaluated via the method of asymptotic decomposition, wherein the singular behaviors are entirely extracted and integrated analytically -- so that the remaining residual components can be computed effectively and accurately via adaptive numerical quadrature. It is also found that, in the spectral domain, the decay of the (numerically-integrated) residual field forms is commensurate to that of their potential-form counterparts, which eliminates the perceived gap between the computation of the field forms and respective potential forms of the Green's functions in the spatial domain. The effectiveness and accuracy of the proposed methodology is evaluated via comparison with relevant examples in the literature. Second, utilizing the derived electromagnetic Green's function for a layered system due to a horizontal electric dipole, the GPR scan can be simulated over a wide range of pavement profiles. Examples are provided for GPR simulation on a three-layer pavement system. By virtue of this forward model, the best match of the GPR scan in terms of the full waveform can be recovered within thousands of simulations via a optimization routine, where the in-situ layer parameters associated with the measurement are found to be equal to the simulation inputs. The accuracy of the interpreted layer thickness from the proposed scheme is verified by ground truth, with average error around 2.3% compared to 7.5% average error for the traditional method. In addition, the proposed scheme allows an evaluation of the relevant pavement properties with no prior assumptions or subjective image adjustments, unlike the traditional method.

Geophysical techniques, such as ground-penetrating radar (GPR), have been successfully used by law enforcement agencies to locate graves and forensic evidence. However, more controlled research is needed to better understand the potential and limitations of this technology in the forensic context. The goal of this study was to determine the potential of GPR using both a 250 MHz and 500 MHz antennae to monitor eight controlled graves with six different burial scenarios using pig carcasses as human proxy cadavers. In addition, a conductivity meter was employed to determine the applicability of using this technology to locate unmarked graves. For the conductivity meter, the data was processed using an EM38 program in conjunction with the SURFER program to display a conductivity contour map of the grid. For the GPR imagery, reflection profile data was processed using the program REFLEXW while horizontal slices were processed using the GPR-SLICE program. Results indicate that the conductivity meter is not a viable option in the detection of clandestine graves when other geophysical tools are available. For the GPR, results indicate that while graves can still be detected after a two-year period, there is a marked decrease in the response, or resolution, of the burial scenarios. Furthermore, burials with grave goods interred along with the carcasses were far more likely to be detected than burials that were interred with no accompanying grave goods. When comparing the performance of the two antennae, the 250 MHz antenna provided increased resolution for large cadavers buried in deep graves.

Ground-penetrating radar (GPR) can be a useful geophysical instrument in the search and detection of clandestine graves in a forensic context. Controlled research in the field of forensic archaeology has demonstrated the applicability of this technology and is vital for improving GPR search methods. The objectives of this research was to evaluate the applicability of GPR, using 250 MHz and 500 MHz antennae, to locate shallow graves containing small pig cadavers in various burial scenarios over a 12 month period. Data was collected on a controlled grid containing six graves at 0.5 m in depth: five graves containing pig carcasses and one control grave. The five graves containing the pig carcasses were devised to test a number of common forensic burial scenarios. The reflection profile data was processed using the computer program REFLEXW. The results demonstrate that the additional grave items did not always increase the detection of the grave for this monitoring period. Further, the low demarcation of the grave containing disturbed backfill illustrated that the hyperbolic reflection features were the result of the pig carcasses and not the disturbed soil. In terms of antenna performance, the 250 MHz data initially provided a higher resolution within the first few months. However, over time the higher detail provided by the 500 MHz data consistently resulted in easily discernable reflections.

The use of geophysical tools to locate clandestine burials involving bodies has seen increasing popularity among forensic personnel. Often, these search methods are important to highlight certain areas where a body may or may not be located prior to utilizing invasive search techniques. Because of the success of these tools within real-life forensic searches, the use of controlled studies that monitor and detect cadavers over certain lengths of time have been increasingly utilized. However, these controlled studies have not monitored various burial scenarios that mimic real-life situations. This study focused on detecting and monitoring six burials containing pig carcasses used as proxies for human bodies and two control burials with a conductivity meter and ground-penetrating radar (GPR) with a 500-MHz and a 250-MHz antenna over a twelve month period. Each burial within this study represented a different forensic scenario that mimicked a real-life situation. Further, forensic use of GPR in both controlled settings as well as real-life searches have mainly focused on the use of a 500-MHz antenna. Therefore, this research also compared the use of a 250-MHz antenna with a 500-MHz antenna. Lastly, a number of GPR imagery options were utilized including reflection profiles and horizontal slices with various GPR software programs to compare the results obtained. Results obtained from the conductivity meter were compared to the results obtained by both antennae of the GPR. Overall, the use of multiple GPR imagery options provided increased resolution of the burial scenarios. Results showed that the conductivity meter was not a beneficial geophysical tool because none of the burial scenarios were detected. On the other hand, the use of GPR showed that the graves with objects added to the pig carcasses provided iv increased resolution compared to the graves containing only pig carcasses. Lastly, the 250-MHz antenna provided better resolution of the burial scenarios than the 500-MHz antenna due to easier discrimination of the forensic targets. Therefore, the use of a 250-MHz antenna would be a viable option to search for clandestine burials containing adult-sized bodies

Analyses of multiple, coincident electrical resistivity and ground penetrating radar surveys successfully imaged a collapse zone and other areas of concern in the vicinity of a previously undocumented, shallow coal mine in Summit County, Ohio. Analyses and inverse modeling of resistivity profiles collected using a Wenner array illustrated that high resistivity regions were associated with collapse of structurally weak sandy shale overlying an excavated coal seam at ~ 10 m depth. Analyses of ~80 radar profiles indicated that such data were more successful at imaging potential near-surface collapse areas than the resistivity method. Futhermore, those data appeared to provide evidence that such collapses propagated from deeper within the section by successive failure of multiple units rather than as a complete structural collapse. The combination of these two techniques illustrated that the spatial extent of actual and potential mine collapse areas can be successfully mapped despite site limitations imposed by property boundaries in an urban setting.

Due to high demand for wide bandwidth applications, UWB antennas have received significant attention in many commercial and military application areas. They can provide very wide bandwidth information with a single antenna configuration. However, designing UWB antennas have very strict requirement such as broadband matching, broad beamwidth, and good efficiency throughout the operational frequency band which is generally difficult to obtain. In this work, the finite different time domain (FDTD) method was selected for the design and optimization of UWB antennas in many different application areas. They include ground penetrating radar (GPR), anechoic chamber feed antenna, near field probe antenna and tapered chamber feed. All these antennas require UWB operation, dual linear polarization, and broad beamwidth. For each application area, they have their own detail operation requirements. With the help of the FDTD code and through understanding, the antennas are deeply studied and analyzed for the final design. This process saves time and cost compared to the repeated prototyping. For the verification of the numerical result, prototype antennas are built, measured and compared to its numerical model result. The measurement and the simulations agree due to the realistic modeling of the geometry.

Ground Penetrating Radar (GPR) is a common technique for locating buried objects in the near surface. The near surface is never perfectly homogeneous due to different moisture levels, grain packing, and types of material that influence the properties in the subsurface. This dissertation examines the influence of heterogeneity on GPR measurements, its influence on spatial dispersion, and defining the GPR response from a range of standard deviations of different numerical models. Most modeling in GPR concentrates on antenna patterns or dispersion caused by complex permittivity in homogeneous blocks of material. The forward model developed in this dissertation incorporates heterogeneity by replacing the traditional homogenous spatial regions with a distribution of physical properties. The models in this dissertation maintain the major spatial model boundaries, but the physical model values within each boundary are determined by a statistical distribution. Statistical approximations of heterogeneity of the physical property distributions can provide an approximation of the geologic noise that influences GPR measurements. This dissertation presents a numerical modeling analysis of random property variation, where the variations occur in one, two, and three directions. The models are developed for a half space and a two layered earth model where the input is a Ricker wavelet. Most of the visible spatial dispersion of the electrical field in both the half space and the layered earth models studied in this dissertation, occurred in the near region of the electromagnetic field. However, the largest average dispersion occurred in the far field at 1.0 m distance from a dipole source. The presence of horizontal layers increased the dispersive effects of the random distribution of electrical property values. There was also a measurable change in the dispersed field when the layers were vertical. There was more change with thin horizontal layers than with tubes or three dimensional variations of heterogeneous material. A practical conclusion of this study is that lateral variation in physical properties must be taken into account when interpreting GPR data.

Ground Penetrating Radar (GPR) is an instrument used in semi-automated construction systems. In principal, images of subsurface objects such as pipes and mines may be detected and potentially measured. The detection of utilities is complicated by a combination of the complexity involved in the data collection technique of the GPR and the irregularities present beneath the surface. This thesis provides the initial results in the development of an algorithm to invert the effects of these corruptions and return images, which are exact in the placement and conformation of subterranean objects. The technique employed is a deconvolution-like method that utilizes a maximum a posteriori (MAP) based optimization method to estimate the best reconstruction. Mean field annealing (MFA) using gradient descent is the optimization method used. Using this technique, single objects in the field of observation were reconstructed to within an acceptable percentage of their original shape. Further work would involve reconstructing multiple objects in the field of observation as well as considering features other than hyperbolae that correspond to objects.

At the turn of the century, the City of Orlando initiated the "Neighborhood Horizon Program." This program involved local citizens to help improve their community resources by engaging in a process of planning where the problems associated with the communities were identified. Many residents favored to bring back the brick roads that were overlaid with asphalt concrete to provided better transportation in the mid 1900s. With majority of the neighborhood streets already bricked, removing asphalt ensured safety, served as a technique for slowing traffic, and added to the historical integrity. Since there were no official documentations available that stated the definite existence of bricks beneath the asphalt surface course, it would have been rather impossible to core hundreds of locations to ensure the whereabouts of these anomalies. Thus, without time delays and excessive coring costs, a nondestructive instrumentation of Ground Penetrating Radar (GPR) was employed in the detection of bricks. This geophysical survey system distinguishes materials based on their different electrical properties that depend upon temperature, density, moisture content and impurities by providing a continuous profile of the subsurface conditions. The Ground Penetrating Radar operates on the principle of the electromagnetic wave (EMW) theory. The main objectives of this study was to investigate the existing pavement by using Ground Penetrating Radar (GPR) in detecting the brick base and to analyze the performance of pavement system for fatigue and rutting. The results of this study will assist the City of Orlando in removing asphalt layer, rebuilding of brick roads, and facilitate in better zoning and planning of the city. The construction of controlled test area provided with a good sense of brick detection, which helped in precise locations bricks for sections of Summerlin Avenue, Church Street and Cherokee Drive. The project demonstrated a good sense of detecting the subsurface anomalies, such as bricks. The validation of the profile readings was near to a 100%.