ABSTRACT
Over the past several years, ground penetrating radar (GPR) systems have been developed and
improved upon to enable the accurate mapping of underground utilities and other structures.
Many companies now consider the use of GPR a standard process in their Subsurface Utility
Engineering (SUE) investigations. Now systems have been introduced that deploy multiple GPR
antennas on one platform to allow creation of 3D images of the subsurface for even more
complete mapping. In addition, these 3D Radar systems have been augmented by multi-sensor
electromagnetic geophysical systems to help image metallic targets. These 3D Underground
Imaging systems help engineers do a better job of mapping whatever is in the subsurface by
producing high data density images of the target zone. Having such data available benefits State
Departments of Transportation, utility owners, and others by aiding in project design and
decreasing project delays due to unknown underground structures that may otherwise be
discovered only during construction. This paper describes such a 3D Underground Imaging
system and then provides relevant case histories.

INTRODUCTION
Thanks to 3D Underground Imaging technology, locating and identifying underground facilities,
especially in large or highly congested areas, has gotten easier recently. This technology lets the
user see what lies beneath the ground in three dimensions for mapping targets such as buried
utility lines, trenches, storage tanks, debris, rock formations and even voids.

When applied to utility investigations, 3D Underground Imaging adds another dimension
to the locating and mapping capabilities of Subsurface Utility Engineering (SUE). In some
situations, especially in areas with few anticipated underground facilities, the more conventional
subsurface utility designating, locating and mapping services are all that is necessary. However,
in areas with a known high volume of underground facilities, such as highly urbanized areas with
multiple buried utility lines, the heavy congestion may cause conventional SUE technologies to
miss some features. 3D Underground Imaging excels in these types of situations because it
"sees" in three dimensions, rather than two, and provides a full view of what lies beneath.

Unlike conventional ground penetrating radar (GPR), 3D Underground Imaging
incorporates a multi-sensor radar system that includes 14 GPR channels instead of just one. It
collects underground data in 5.12-ft-wide swaths that are merged together to create full 3D
images of the underground.

In addition to 3D Radar, 3D Underground Imaging also includes electromagnetic
technology in the form of a multi-sensor Electromagnetic Induction (EMI) system, which allows
for the rapid and accurate detection of conductive facilities and objects. EMI complements the
3D Radar data to enhance the overall result of a 3D Underground Imaging investigation.
3D Underground Imaging has proven to be an attractive option for characterizing the
subsurface in large or complex areas.

MEASUREMENT SYSTEMS
3D Underground Imaging technology consists of a multiple sensor 3D Radar system, a multiple
sensor Electromagnetic Induction (EMI) system, and state-of-the-art geophysical 3D data
processing, interpretation and management software.

Multiple Sensor 3D Radar System
The 3D Underground Imaging 3D Radar system is a 14-channel, cart-based ground penetrating
radar unit that is typically towed behind a vehicle during survey operations as shown in Figure 1.

FIGURE 1 3D Radar system with dedicated tow vehicle

The GPR system can acquire data at speeds up to 5 miles per hour (8 km per hour) across
the ground surface with acquisition timing controlled by a wheel encoder. Positioning
information is logged with a global positioning system (GPS) receiver or robotic tracking
theodolite.

3D Radar data are post-processed using standard GPR algorithms such as deconvolution,
frequency filtering, background removal and 3D Kirchhoff migration. Processed data are then
merged with the appropriate survey data files. The multiple survey swaths, geo-referenced to the
desired coordinate system, are assembled into a composite 3D data block of the project area from
which buried utilities and other subsurface features are interpreted.

It should be noted, however, that although differences between conventional singlechannel
GPR and Multi-channel 3D GPR are numerous, the underlying physical laws governing
the propagation of electromagnetic energy pertain to both methods equally. Namely,
investigation depths of 3D GPR are limited, in the same manner as conventional GPR depths are,
by electrical properties of soils. Highly electrically-resistive soils (dry sands) allow for the
investigation of the subsurface to depths exceeding 20 ft (6.1 m), while highly conductive soils
(wet clays) may prohibit penetration to depths greater than 3-4 ft (0.91-1.22 m).

Multiple Sensor EMI System
The 3D Underground Imaging multiple sensor EMI system is a Time Domain Electromagnetic
(TDEM) system consisting of three pairs of transmitter and receiver coils configured and
synchronized to operate simultaneously as a single array along a survey swath. The TDEM coils
are mounted on an electrically nonconductive deployment cart that is towed behind a vehicle as
shown in Figure 2. The complete system is able to cover a five foot wide swath in a single pass at survey speeds up to 5 miles per hour (8 km per hour).

The coil systems are time domain-based metal detectors which detect both ferrous and non-ferrous
metallic objects and other changes in electrical conductivity.

FIGURE 2 Multi-sensor EMI system with
positioning hardware

3D UNDERGROUND IMAGING CASE STUDIES

Case Study #1 – 3D Radar, Power Plant Expansion, North Carolina

Introduction
A North Carolina fossil plant was faced with enacting a flue gas desulphurization retrofit
program in order to comply with the latest environmental regulations. The environmental retrofit
program for the 70 year-old plant was to include significant construction and excavation at the
existing plant site to include the installation of several 36 in (91 cm) diameter concrete caissons
to support new structures. During the early design stages of the project, the design engineers
wanted to account for all known and unknown subsurface utilities as well as potential unknown
underground obstacles that would negatively impact the costs and timetable of design and
construction activities. Due to the age and complexity of the facility, the designers assigned high
probability to the potential presence of subsurface utilities that may have been represented
incorrectly on existing maps, or omitted entirely. The designers also decided that the advanced
age of the plant would increase the probability that construction activities occurring on the site
since the 1930’s, resulting in buried underground obstacles (structures, debris), may have gone
unrecorded.

Geophysical Investigation
Due to suspicions of the presence of buried underground obstacles, in addition to an array of
known and unknown subsurface utilities, the designers opted to employ 3D Underground
Imaging on their project, in addition to a complete SUE investigation. Considering the unknown
nature of materials composing potential subsurface features, the 3D Radar technology was
chosen for its ability to image all materials, regardless of electrical conductivity characteristics.
The project designers selected areas for investigation totaling approximately two acres
where construction activities had the potential to come into conflict with subsurface features.
The study area was then subdivided into three subsections based on site logistics and data file
manageability considerations. Over a three day period, virtually every accessible square foot of
the study area was covered with the 3D Radar system, resulting in 81 3D Radar data swaths. In
comparison, in order to achieve similar data densities through a conventional single-channel
GPR investigation, more than 1100 GPR transects would have been required.

Real-time positioning of the 3D Radar antenna arrays during data collection, as well as
the accurate positioning of corresponding subsurface features interpreted through data postprocessingand analysis, was acquired through the use of a differentially-corrected GPS (DGPS) system, achieving sub-decimeter (3.93 in) horizontal accuracy. Following field data collection activities, the GPR data were processed with RADAN Version 6.0 from Geophysical Survey Systems, Inc. (GSSI) using standard algorithms. The processing filters applied to all data were background removal, range gain adjustment, deconvolution to remove multiple reflections, stacking, and band-pass filtering. Through this digital filtering process, near surface features and deeper targets become more readily apparent in the data set. Velocity analysis was also performed on several of the GPR data sets to better determine the dielectric constant of the subsurface soils and ultimately a more accurate estimate of the target depths. After processing the individual data files, the GPR swaths were assembled into several 3D Radar data blocks representing the surveyed areas. Interpretation of the 3D data set allows slices to be made down through the data block with respect to time (depth). In this manner, the linear trends representing buried utilities and other features may be illuminated to gain an understanding of their shape, amplitude, depth, location and orientation. Interpreted target locations and depths were determined directly from the 3D data set, recorded, and transferred to project drawing files.

Results
Analysis of 3D Radar data blocks yielded the identification of approximately 120 distinct
subsurface features. The majority of these features coincided with utility location information
garnered through the SUE investigation and/or plant utility plans. However, several additional
previously unknown targets were identified as well. Figure 3 illustrates the phenomenon through
which features are illuminated and interpreted at varying depths within each 3D Radar data
block.

FIGURE 3 Series of depth slices from a 3D Radar data block to illustrate how targets appear and
disappear as slices of varying depth are viewed. (A) Plan view section of GPR data taken at a slice 8 inches below ground surface in. At this shallow depth, the only features that can be identified are the surface railroad tracks and ties. (B) Depth slice taken at 15 inches depth. The railroad tracks are less distinct, but a feature interpreted as a drain and catch basin are now visible. (C) Depth slice at a depth of 33 inches. In this view, deeper linear features that may correspond to buried utilities can be seen.

Of significant importance to the designers of the plant expansion was the identification of several unknown utilities and areas of buried reinforced concrete, buried rebar, and buried railroad ties that coincided with the preliminary design locations of several of the 36 in (91 cm) concrete caissons. After receiving CADD drawings with accurately positioned subsurface
features interpreted from the 3D Radar data, the designers were able to decide in which areas
they would amend their initial design based on the new subsurface information, and in which
areas they would maintain their initial design and remove the subsurface debris at select conflict
points. In either scenario, the 3D Radar information aided the designers in a decision-making
process that would ultimately increase the probability of avoiding unexpected redesigns and/or
significant construction delays.

Case Study #2 – 3D Radar, Sewer Line Elevation Profiling, Florida

Introduction
A local engineer consulting with a Florida city was tasked with the rehabilitation of a 16-inch
diameter ductile iron sanitary force main that was experiencing significant corrosion and
subsequent failure at certain locations throughout a 1000 ft (305 m) length of pipeline. The
engineer and city officials had reason to believe that localized high points in the pipe caused by
sudden changes in the pipe elevation (slope) were trapping sewer gases against the pipe wall in
small pockets. Over time, the gases within these pockets were interacting with the pipe wall and
causing accelerated corrosion, which when left unchecked, was contributing to localized failure
of the pipeline. In order to prevent future pipe failures related to this phenomenon, the “high
points” needed to be identified and these sections of pipe replaced with straight sections.

Geophysical Investigation
3D Underground Imaging was chosen to be employed on this project due to the ability of the 3D
Radar system to cover a 5.12 ft (1.56 m) wide swath in one pass, as well as its capability of
tracking the position of the radar system (and subsequent GPR data) to survey-grade tolerances
in both the horizontal and vertical dimensions.

FIGURE 4 3D data screen that was analyzed as part
of the utility elevation profiling.

Field data collection consisted of a single 3D Radar data swath centered over the approximate horizontal position of the force main, requiring less than one
day to complete. The real-time position of the 3D Radar system was tracked and recorded automatically during data collection using a robotic survey total station. This enabled each of the 168,000 GPR data points along this 1,000 ft (305 m) data swath to be positioned to surveygrade
accuracy on the local survey datum provided by the city.

Results
By tracking a survey prism mounted at a fixed point on the 3D Radar system, the robotic total
station was able to produce an accurate survey (elevation profile) of the ground surface over the
entire length of the 3D Radar data swath. In order to calculate the depth to the force main, a
radar velocity analysis was performed and a representative dielectric constant was assigned to
the subsurface soils.

Depth values (below ground surface) of the force main were then converted to elevations
by comparing them to the elevation profile of the ground surface. A profile plot of elevations of
the force main was then produced, allowing the client to analyze the actual profile of the force
main for sudden slope changes.

Following analysis of the force main elevation profile, those sections of pipe
experiencing sudden slope changes were selected for further ground-truthing and eventual
rehabilitation.

Case Study #3 – 3D Radar/EMI, Airport Expansion, New Jersey

Introduction
In order to extend runways as part of a minor airport expansion, project designers required
utilities beneath an adjacent roadway to be accurately mapped for their eventual removal or
replacement during construction efforts. The locations of at least four utilities were known to
exist in the affected area.

Geophysical Investigation
In addition to a standard Subsurface Utility Engineering investigation, the designers opted to
employ 3D Underground Imaging to supply them with the most comprehensive and accurate
subsurface information available.

FIGURE 5 EMI contour map overlaying site aerial imagery that illustrates a strongly conductive linear feature (EM1) interpreted to be a gas main and a weakly conductive unknown crossing feature (EM2). The two white ovals represent passing automobiles.

3D Radar and EMI survey swaths were conducted along the approximately 1,000 ft (305
m) section along the road to be moved as well as along a walking path adjacent to the roadway.
In the roadway, the ability to collect data at significant speeds allowed for data swaths to be
acquired with the flow of traffic in the appropriate travel lanes. This resulted in minor disruption
of normal traffic patterns as travel lanes were not required to be shut during data collection. The
3D Underground Imaging survey resulted in approximately one acre of coverage within the study area. It is noted that the amount of data acquired with the 3D Radar system during the four hours required to perform the study is equivalent to 336 individual GPR transects with a conventional singlechannel GPR system

Following data collection, both sets of 3D Underground Imaging data were interfaced with DGPS data and adjusted to appropriate state plane coordinates.

Results
Results of the 3D Underground Imaging investigation generally agreed with available published
utility information and SUE results. The lone feature discovered solely through the 3D Underground Imaging investigation was a small diameter utility crossing the roadway, as evident in the EMI contour map shown in Figure 5. Interestingly, this feature was not accounted for in the 3D Radar dataset, indicating that the feature was either too small to be imaged or was located at a depth greater than the maximum radar penetration depth. Project designers were able to use the utility information in the resulting CADD utility plans to proceed confidently in their overall project design.

CONCLUSION
3D Underground Imaging represents a major evolution in the application of geophysical
technologies for the underground mapping of utilities and structures.

3D Underground Imaging employs a 14-channel 3D Radar system, a multi-sensor
Electromagnetic Induction system, accurate real-time positioning systems, and leading edge
geophysical data management software to collect high density geophysical data and create three
dimensional depictions of subsurface environments which are more accurate and comprehensive
than those created with traditional two dimensional technologies.

Whether employed as a complement to a comprehensive Subsurface Utility Engineering
investigation, or utilized as a stand-alone method to image other subsurface objects or features in
large or complex areas, 3D Underground Imaging is proving to be a valuable tool in the quest for
subsurface characterization.

The application of 3D Underground Imaging to the pre-design phase of a project can
enhance the accuracy of project designs and cost estimates and help to streamline construction by
using all three dimensions to capture as much of the total underground picture as possible.

Credits
Author(s)
Christopher R. Proulx
Gary N. Young, P.G.

Submitted To:
Transportation Research Board
Utilities Committee