When working with GPR data, specialists may
encounter a situation where the quality of the result of the GPR profiling
performed by them is worse than in the promotional materials of geophysical
equipment manufacturers. The GPR advertising profiles clearly show the
boundaries between the layers and diffracted reflections from local objects,
from which it is easy to determine the wave propagation velocity in the
subsurface environment and the permittivity in the layers. However, this does
not mean that GPR manufacturers are trying to mislead potential customers. It's
just that these exemplary GPR data were obtained during GPR profiling of
subsurface with low losses and contrast layers that best reflect the quality of
GPR equipment.
The
following is an example of a GPR profile obtained during the investigation of a
high-contrast subsurface medium with low losses. This field record does not
require processing, since it already contains enough information about the
structure of the subsurface.
But along with such subsurface environments favorable for GPR surveys it is necessary to investigate low-contrast strata with a high level of absorption of electromagnetic energy, which do not have sharp transitions between layers, under conditions of interference of various nature. Often there are such field recordings:
This profile is recorded by GPR 200 MHz on the sea beach. During the profiling process, the GPR moves away from the coastline in a direction perpendicular to it. The peculiarity of this survey is that the coastal saline soil is probed, which has a high conductivity. For this reason, GPR pulses quickly fade out, and already at a shallow depth, the intensity of reflections from the boundaries of the layers becomes comparable to the noise level on the GPR profile. Noise and interference mask useful reflections, so visual analysis of this profile does not give a positive result. The use of various types of filtering to remove interference also does not lead to success. After removing the interference, the boundaries of the layers are still not visible:
The fact that after suppression of
interference on the GPR profile, reflections from the boundaries of the layers
do not become noticeable does not mean that they are absent. When useful
reflections are close in their characteristics to interference, during the
filtering process, these reflections are removed together with interference.
The BSEF automated analysis algorithm works more selectively, and is able to
separate useful signals from interference close to them.
The
figure below shows a section of the Resistivity attribute created based on the results of automated BSEF analysis of the GPR
profile in question. In contrast to the visually uninformative profile, the
attribute section displays the structure of the subsurface environment well.
Within the investigated subsurface, two main layers are distinguished. The
upper layer is characterized by reduced resistivity values, this layer is
displayed mainly in blue colors on the section. The lower layer has higher
resistivity values and is represented by yellow-red shades. According to the
section, it is possible to trace how, as you move away from the coastline, the thickness
of the layers of coastal soil and the electrical resistivity inside these
layers change.
A visual representation of the variability of the section attribute can be obtained using the statistical module of the GEORADAR-EXPERT software system, which provides information in the form of graphs and tables for 12 statistical indicators. Indicators can be calculated for each layer, for user-specified layers, for each layer boundary, or for the entire section as a whole. As an example, the figure below shows the result of using a statistical module in the form of a graph of changes in the average values of the electrical resistivity in the upper layer of the Resistivity attribute section. The area of the graph that exceeds the user-defined threshold is highlighted in red. At the bottom right is a thumbnail of the section, which shows the boundaries of the layers. The fill highlights the layer whose data is represented by the graph.
It is obvious that the switch from the
representation of data on the subsurface environment in the form of a set of
amplitudes of reflected signals (GPR profile) to the characteristics of this
environment obtained as a result of the application of the BSEF automated
analysis method (attribute section) significantly increases the informativeness
of the GPR survey. The result, presented in the form of an attribute section,
is more understandable to the customer of this survey. The list of attributes
used in the GEORADAR-EXPERT software system is quite wide. The list of attributes and their description can
be found in the Section List of Attributes section of the user manual, which is
available for download at the FOLLOWING LINK.
One of the advantages of automated BSEF
analysis over other methods of processing GPR data is the ability to study
subsurface environment with this method, the electrophysical characteristics of
which change vertically smoothly, without sudden jumps. Abrupt changes in the
properties of the subsurface in the vertical direction create conditions for
the reflection of probing pulses and their reception by GPR. Environments that
have a smooth character of changing electrophysical properties cannot form
reflections. In such places, only high-frequency noise and various kinds of
interference are present on the GPR profile, provided that the studied
environment in these places does not contain local objects.
Local objects are objects in a subsurface
environment whose linear dimensions are comparable to the wavelength of the GPR
probing pulse, and the electrophysical characteristics of these objects differ
from the electrophysical characteristics of their host environment. For
example, stones in the ground can act as such objects. Local objects have an
important feature. It consists in the fact that when the GPR probing pulse
interacts with local objects, these objects become a source of diffracted
waves, the kinematic and dynamic characteristics of which carry information
about the properties of the subsurface environment.
Due to the low intensity of diffracted
waves, which is often comparable to the level of interference, visual analysis
can detect only a small part of these waves on the GPR profile. As a rule, this
amount is not enough to obtain a detailed information about the structure of
the subsurface environment based on the measured characteristics of these
waves.
Further, the figure on the left shows the
result of GPR profiling of soil with a smooth change in electrophysical
characteristics. The profile was obtained by a GPR 250 MHz. The profile crosses
the dry bed of a seasonal stream, the talweg of which is 25 m from the
beginning of the profile. The electrophysical characteristics of soils in this
profile change smoothly with depth, without sudden jumps forming reflections on
the GPR profile. For this reason, the profile does not contain characteristic
extended reflections, which are interpreted as reflections from the boundaries
of the layers. Under these conditions, it is possible to obtain information
about the structure of the studied environment using automated BSEF analysis.
The
figure on the below shows a section of the Resistivity attribute. The section of the attribute shows the position and shape of the
channel stream deposits, as well as the distribution of resistivity within
these deposits. The section shows that the boundary between dense soil and
loose sediments has a concave shape and reaches a maximum depth of 5.5 meters
in the talweg area of the stream. The lowered values of the section attribute
indicate that the stream has not completely dried up, and the lowest
resistivity, indicating the highest humidity, is located in the talweg area of
the stream, at depths of more than 1 meter.
The processing time spent on performing
automated BSEF analysis and section calculation is relatively small. Processing
of the GPR profile in question lasted a little less than a minute on a computer
with a quad-core 2.4 MHz processor and Windows 10 operating system. Significant
time savings occur in the case of processing a large amount of GPR data in
batch mode.
Saving time is especially relevant for
processing the results of GPR profiling of extended multi-kilometer objects,
such as roads and railways, or the threads of trunk pipelines. After setting up
the parameters and starting the automated processing process, the user can
switch to other tasks, and the software system will independently load GPR profiles
and save the processing results to the computer's hard disk.
The GEORADAR EXPERT provides for the export
of processing results to graphic formats, to MS EXCEL spreadsheet formats, to
ASCII universal text format tables, as well as to the GRD grid data format of
the Surfer software. Exporting data to these formats makes it possible to use
the results of processing more widely using third-party software. For example,
for additional analysis or integration into various geoinformation systems.
Images of attribute sections or 3D assembly slices saved in a graphical format
can be used as a substrate in a geo-base drawing in AutoCAD or for insertion
into a GPR research report.
Sections Summation
The GEORADAR-EXPERT software system has a
fairly wide set of attributes for solving a whole range of GPR tasks. It
happens that a section of one attribute does not provide complete information
about the object under study, but various fragments of this information are
distributed across several attribute sections. In this case, the summation of
sections allows you to combine disparate information about the object of study
into one summary section. Also, summation eliminates artifacts caused by the
accumulation of errors in the process of collecting and processing GPR
information.
The
figure below on the left shows the GPR profile recorded during the study of the
paleodoline by a GPR 100 MHz. The center shows a set of attribute sections for
summation, created based on the results of automated BSEF analysis of this
profile. The result of summing these sections is shown on the right.
Visual analysis of the GPR profile shows
that this profile is uninformative, starting from a depth of 2 meters. The base
of the paleodoline, according to a priori information, is below this value. The
attribute sections show some details of the buried relief, but each of the
sections separately does not provide complete information about the structure
of the studied subsurface.
The
summation operation combined this disparate information into one whole. The
relief of the base of the paleodoline on the summation section is well traced
along its entire length. In the figure, the base of the paleodoline is
indicated by a dotted line. This example illustrates how the use of the
summation module allowed solving the problem of insufficient information
content of the GPR profile and individual sections of attributes by this
profile.
Automated BSEF Analysis Advantages
The use of automated BSEF analysis has the
following advantages over other methods of processing GPR data: