Professional Surveyor March 2008 Volume 28 Number 3
Feature: Surveying Goes Underground
John Krause and Andrew Lund

For as long as there have been surveyors and underground utilities, surveyors have been asked to make representations on maps about the lines. Without access to pipe inverts at manholes, they are understandably reluctant to do so. "How can you survey something you can't see?" "You expect me to certify or attest to conditions based on assumptions from records, surface features, and paint markings by others beyond my control?" Often, the surveyor accepts liability to get the project or puts disclaimers on the map the client won't like. Relying on one-call services to depict horizontal locations of buried utilities on maps equates to Russian roulette unless you have a deep understanding of locating equipment. Another problem: You never really know whose hands your maps will end up in, or if the person will read all the notes. Even when you win, litigation is stressful and costs time and money.

But this roulette game may be ending. Thanks to new technologies, it has become possible to survey the unseen underground, and this technology has fallen under the oversight and quality control of the professional surveyor. It's not entirely accurate to call these technologies new. They have, in fact, been around for a while. But not until recently did they assume a form suited to the underground challenges faced by engineers, surveyors, and contractors.

It actually goes back over 40 years … to a college dorm room at the University of Rochester. Not so much the room, but the roommates: Alan Witten and Tony Devaney. They weren't your typical college students; both were considered top pupils of Dr. Emil Wolf, who had just published the (still) seminal text in the field of optics, The Principles of Optics (Macmillan, 1964).
Twenty years later, (now Dr.) Devaney would patent a concept called geophysical diffraction tomography (GDT). Stated simply, GDT involves transmitting a signal through the earth (or water), bouncing it off objects contained therein, receiving the signal with an array of antennas, and processing the (now multiple) signals with complex algorithms. The signals return to each antenna in the array nanoseconds apart, with the slight time difference allowing for 3D triangulation and ultimately visualization of the objects the signal(s) bounced off of. Originally, GDT was applied in acoustic form (signal equals sound).

At the time, Devaney was working for the French petrochemical giant Schlumberger, so exploring for oil was a natural fit as GDT's first commercial use. It is still used for that purpose, albeit on a much grander scale. While Devaney was using GDT to find new oil reserves, Dr. Witten was doing more "fanciful" (read: interesting) things. From visualizing ancient underground cave villages in the Negev desert in Israel to detecting tunnels in the Korean demilitarized zone to imaging sunken galleons off the coast of Madagascar, Dr. Witten put GDT through its paces. Perhaps his most notable application of GDT was in the New Mexico desert in 1989, where he helped unearth Seismosaurus (the "Earth Shaker"), the largest dinosaur on record. "Sam," as he was later nicknamed, would become the logo of the company Witten would ultimately found.

Chance Encounter
This brings us closer to real time. It's 1992, and Dr. Witten is working near Jacksonville, Florida for Oak Ridge National Laboratories, trying to image submerged pipes. By sheer coincidence, he encountered Robert Green, a local utility contractor. As a second-generation contractor, Green was painfully familiar with the problems that arose from not knowing what's down there. He asked Dr. Witten what would prove a pivotal question: "How can we make this geophysical telescope (acoustic GDT) into a geophysical microscope?" When Witten replied, "GPR," proposing to use ground-penetrating radar as the signal, the concept of computer assisted radar tomography (CART) was born.

Soon thereafter, in 1994, the two men formed Witten Technologies in Boston, MA to commercialize GPR-driven GDT. It took the company the remainder of the century to do so, with CART, shortened to RT (radar tomography) for simplicity, debuting around 2001. Along the way, the company received assistance from the Electric Power Research Institute, the Gas Technology Institute, and Consolidated Edison. In fact, ConEd was Witten's first true client, and the first notable RT work was done in lower Manhattan. In a strange twist of fortune, the company scanned the streets around the World Trade Center just prior to 9/11 and then helped ConEd sort through the subsurface spaghetti left in the tragedy's wake.

Since then, the technology has been applied to over 20 million square feet in the United States and abroad.

Since GPR is RT's foundation, an understanding of RT requires some basic GPR knowledge. Featured numerous times in Professional Surveyor Magazine, a typical GPR system involves a single pair of antenna—transmitter and receiver—mounted in a pushcart or attached to a vehicle. The transmitter sends an electromagnetic pulse downward, which bounces off objects and features and returns to the receiver. The curves and ripples in the resulting image(s), once interpreted, reveal the 2D position of certain objects and features. While it has been used effectively in a variety of applications, the inherent subjective interpretation requirement lends it an air of "rocket science." RT has taken GPR from 2D to 3D, eliminating much of this interpretation requirement.

For ease of explanation, radar tomography can be broken down into three separate but interrelated components: hardware, process, and software.

Hardware:
An array (in both 200- and 400-MHz versions) consists of 17 ultra-wideband GPR antennae (9 transmitters and 8 receivers) and a control box that manages the firing of the antennae in a controlled, rapid sequence. The firing sequence can be either timed or triggered by a survey wheel mounted on the array. A multiple array with sufficient distance between the outermost antennas is key to creation of 3D data. The array is contained in a shielded, eight-inch-thick, pool-table-sized box, either front-mounted on a commercial mower chassis or pulled behind a vehicle. Mounted on the array is an adjustable prism pole with a 360-degree prism at the top. A robotic total station tracks the prism (ergo, the array and all the antennas contained therein) at all times in all three planes.

Process:
As the array passes over the ground at about two MPH, it is tracked by a constant laser connection between the total station and prism pole. In addition, the total station and prism pole are used to auto-collect numerous control points. In addition to providing the required degree of geospatial accuracy, these points are also used to stitch passes together to create seamless coverage over large areas. Thousands of multi- aspect (that is, the same spot from multiple angles) 2D GPR images are created for a given project or area. Using a combination of automated and manual software processes, this 2D data is later converted into a 3D data cube.

Software:
The image-processing algorithms act as a mathematical lens, which, much like the lenses in our own eyes, focus indistinguishable optical patterns into recognizable images. In the first phase, the software merges the radar and positioning data to focus the 2D radar echoes into 3D/motion imagery. In the second phase, additional software then analyzes the data cube to create plan-and-profile drawings of all detected lines and features in the standard CAD format(s) used in engineering design and construction.

Florida Shows Early Interest
Besides ConEd in New York, a few other organizations from Witten's home state of Florida took an early interest in the technology. Miami-Dade Water Sewer used RT extensively, ultimately issuing an RFP specifically for radar tomography services. The Florida Department of Transportation, always in search of cost-cutting measures for an ever-expanding slate of highway improvement projects, engaged in an exhaustive three-year pilot study. Sunshine State One Call of Florida, the state's call-before- you-dig center, participated in the study.

The lessons learned from the FDOT pilot study, performed across four different projects totaling roughly two million s/f², were many and varied.

At first, it was a lesson in communication. Witten consisted of scientists who knew they had a good thing but didn't know how to deliver it. FDOT really didn't know what to expect, and Witten didn't know what FDOT expected … kind of like two people who didn't know how to dance learning to dance together. Frustrations ran high on both sides, and the department was close to giving up. Witten hired John Krause, PSM, a 15-year veteran of subsurface utility engineering (SUE), and enlisted the aid of Craig A. Smith and Associates (CAS), a civil engineering, surveying, and SUE firm out of Fort Lauderdale.

Witten had always understood the importance of this data because it provides the user with accurate and contiguous x/ y/z's for the tops of all the lines and objects it "sees." FDOT realized this, too. The disconnect was that neither had any experience blending RT with existing survey, engineering, and SUE practices and processes, then presenting the data in formats familiar to engineers, surveyors, and contractors. In other words, the deliverable format did not meet industry standards and statutory mapping requirements, and neither party knew how to fix it.

With Krause and CAS came this knowledge, ultimately righting the FDOT ship. CAS would become Witten's first licensee. Further complicating the above issue was the fact that FDOT was introducing MicroStation Version 8 about the same time. When the coordinate richness (too many unique x, y, z coordinates) of the RT data crashed the system, CAS' and FDOT's CAD teams worked together to come up with a solution.

The study produced estimates of two-thirds less vacuum excavation and a 10-to-1 return on investment, and two overarching themes became apparent in the long run. First, RT was not the magic bullet some perceived it to be. (This was also an expectations-management lesson for Witten.) While it is a great tool that makes for a more effective toolbox, it doesn't replace any other tools in the box. Like GPR, it can't see everything under all conditions; certain very small lines (i.e. fiber optic) are hard to detect, and certain soil conditions (i.e. high conductivity) obscure the signals. Second, it's only applicable for specific kinds of projects—large, urban ones with lots of underground utilities in the way—so project selection is crucial. Ultimately, FDOT would include RT in its roadway design "bible," the 2007 Plans and Preparations Manual.

Another System Comes Along
Along the way, Witten's scientists recognized some of RT's shortcomings, including limited penetration depth and dependence on soil condition, and developed a "sister" system. While it is based on similar arrayed sensing principles, the Arrayed Inductive Receiver (AIR) system is based on conventional electromagnetic (EM) induction locating principles, where an EM sensor detects the position of the utility lines by detecting the electromagnetic fields created by currents imposed on those lines. Conventional EM-based locating consists of two basic components, a current-inducing mechanism to energize a given utility line and a means of interpreting the behavior of that current with a single EM receiver on the surface to provide horizontal location data for that line. Much like RT is an extension of GPR, the AIR system is an extension of EM.

The AIR system consists of several components. At the heart of it is the array itself, which consists of 16 specially designed triaxial broadband electromagnetic sensors. The broadband sensors allow the recording of many different frequencies simultaneously, allowing the operator to use multiple transmitters for detection of multiple utilities. Equally important is an electrical current source, either direct or indirect. The system employs off-the-shelf EM transmitters for direct current application but uses a specially designed transmitter ball for indirect application.
To capture the information coming from the sensors, a 48- channel (6 receivers x 3 channels and x, y, and z for each) data acquisition system is employed. For positioning, the same total station used by RT is used. Most of the components are housed in the AIR trailer, connected to a laptop in the tow vehicle that records the sensor data. And of course, there's the litany of back-end processing components … data processing, visualization, interpretation, and CAD/GIS mapping software.

In a multi-step software process, the data recorded by the sensors is pre-processed, filtered, and merged with the positioning data to create color imagery that depicts the electromagnetic fields. By modeling the resulting data, precise positions for detected utilities are derived. As with the RT system, the final deliverables of the AIR system include both imagery and underground CAD maps of the located pipes.

Recognizing the ability of both systems to solve the problems posed to the transportation industry and other stakeholders (including the general public) by lack of accurate subsurface utility maps, the U.S. DOT's Pipeline and Hazardous Materials Safety Administration selected Witten Technologies to create a merged-array system. The project, informally known as the Dual Array Project, had the title Digital Mapping of Buried Pipelines with a Dual Array System. The stated goal: "develop a non-invasive system for detecting, mapping, and inspecting ferrous and plastic pipelines in place using technology that combines and interprets measurements from ground penetrating radar and electromagnetic induction sensors." The purpose of this research was to demonstrate that the RT and AIR systems would provide complementary information that could be merged to create better and more accurate utility maps for both conductive and non-conductive facilities over large areas.

The capabilities of these two systems, together or separately, change the subsurface liability landscape, removing much of the risk and cost typically associated with "underground ignorance," the bad luck to find objects long abandoned from activities long forgotten on (under) land your client just purchased. The paradigm is shifting from "We'll deal with it when we hit it" to "Let's figure out what's down there before we get started." Those that work underground are starting to realize it's cheaper and more efficient to deal with it ahead of time.
Witten's technologies are just scratching the surface. Having focused almost exclusively on the utilities and transportation sectors since birth, it is just starting to dabble in other (more interesting) applications where seeing the unseen is also of great benefit such as oil and gas, archeology, forensics, environmental, military, etc. The sky … er … the earth is the limit.

About the Authors
John Krause, PSM is vice president of operations for Witten Technologies and has more than 35 years of land surveying experience, 21 as a licensed land surveyor. As a principal of the company, his responsibilities include executive management of the contracts and operations associated with Witten's land surveying and subsurface utility engineering (SUE) projects as well as its licensee program. For the last 13 years, Mr. Krause has devoted his career exclusively to SUE. He can be reached at j.krause@wittentech.com.

Andrew Lund is business development manager at Witten with over 20 years of technical marketing experience, the last 10 focused on the underground damage prevention and SUE industries. He is responsible for marketing, sales, promotions, outreach, media/public relations, and licensee support. He can be reached at a.lund@wittentech.com.