Geoscience Reference
In-Depth Information
7.5.2 i
in t e R P R e t a t i of in
GPR can be used to locate any object that has an electrical properties contrast with the surround-
ing ground, is within the detection range of the radar waves, and is not masked by clutter or noise.
Scattering reflections are caused by an abrupt change in the electrical properties (primarily electric
permittivity) in the subsurface. Some common features that have a high contrast include empty cavi-
ties, voids, or tunnels; changes in rock porosity; the water table; metal objects (e.g., barrels, tanks,
pipes, etc.); plastic containers; concrete foundations; oil, petroleum, dense nonaqueous phase liquid
(DNAPL) spills; or changes in geology.
Interpretation is the intellectual (human or intelligent computer) process of identifying anoma-
lies on the GPR data and determining the nature (size, shape, and physical properties) of the object
in the subsurface that is causing each anomaly. A good interpretation is the result of the skill of the
interpreter (or sophistication of the pattern recognition algorithms), the quality of the data recorded
in the field, and the clarity of the processed display used for interpretation. The interpretation begins
with a good display that makes it easy to identify anomalies, with interpretation and processing
inevitably overlapping each other. Data processed to the point where ready for interpretation should
contain a minimum amount of noise (either random noise or coherent noise). Coherent noise can
consist of features that are a part of the system (e.g., antenna ringing) or objects that are not a target
of the survey (e.g., geologic features, overhead cultural features, etc.). The objects that are not a
target of the survey are often called clutter. GPR data interpretation should progress along the fol-
lowing stages, with some overlap and feedback between stages:
1. Optimize the two-dimensional display to isolate the distinctive anomalies in the data.
2. Identify and classify the anomalies on the two-dimensional displays. Isolate the anomalies
of interest from the clutter in the data.
3. Formulate the three-dimensional display and optimize the display to isolate the trends in
the data.
4. Plot out time slices and cross sections of the three-dimensional data display to provide a
final interpretative view of the anomalies.
5. Classify and identify features on the two-dimensional and three-dimensional data displays.
6. Map the map location and depth of identified objects.
Reflection anomalies on GPR records caused by linear objects (e.g., fences, overhead pow-
erlines, corners of buildings, etc.) located above the surface of the ground are easy to identify
by calculating the velocity from the reflection hyperbolas. The velocity of a radar wave in air is
approximately 1 ft/ns in English length units (0.305 m/ns in metric units), and the velocity of elec-
tromagnetic waves through all other materials is much slower than the velocity through air. Linear
surfaces (e.g., parallel fences, walls of buildings, overhead pipes, etc.) are more difficult to identify
directly from the data and may require the use of good field notes in order to identify them.
Noise from external radio wave sources can be identified by the fact that it is usually semicon-
tinuous and tends to contaminate a series of traces. Radio wave, or microwave, frequency noise on
GPR records can sometimes be minimized by digital signal processing, if the frequency of the noise
is outside the operating frequency range of the antennas.
