Automated BSEF Analysis

Professionals using ground penetrating radars may encounter a situation where the quality of the results of GPR profiling differs for the worse from the samples presented in advertising materials of manufacturers of geophysical equipment. On such advertising radarograms, one can easily find the boundaries between the layers, as well as diffracted reflections from local objects, by which it is easy to determine the speed of propagation of waves in a subsurface subsurface environment and permittivity. As a rule, such results can be obtained during the GPR study of subsurface with low losses and the presence of sharp changes in the electrophysical characteristics in the contact zone of the layers. In practice, one has to study not only low-loss subsurface, but also subsurface environment that have high absorption of electromagnetic energy, without sharp boundaries, with various interferences. And as a result of such a field study, there may be something like this radarogram:

The above example is the result of GPR profiling of bulk soil with a 200 MHz antenna. Due to the strong attenuation of the GPR probe pulses in this bulk soil, weak reflections from the bottom of the embankment are masked by more intense interference. Application to this GPR profile of standard methods of digital signal processing, such as subtracting average GPR traces and frequency filtering in order to suppress noise, did not lead to a positive result. The boundary between bulk and bedrock is still not visible. There are also no diffracted reflections from local objects in the soil, from which it would be possible to determine the wave velocities and permittivity in different parts of the GPR profile in order to rely on these values, at least roughly calculate the position of the boundary between bulk and primary soils. Visual analysis of the processed GPR profile gives practically no useful information:

The following shows the result of processing this GPR profile using BSEF automated analysis in the form of a section of the attribute "The real part of the complex relative permittivity" This is the most commonly used attribute. Undoubtedly, the information content of the result is significantly higher than that of the radarogram. The section can be easily divided into two layers. The top layer is bulk soil. On a section, this layer has colors ranging from light blue to dark red. The bedrock is represented by shades of blue.

By studying the change in the color gamut of the section, one can learn about changes in the electro physical characteristics of the soil inside each layer. For a more accurate representation of changes in the attribute of a section, you can use the statistical module implemented in the GEORADAR-EXPERT software and obtain data for each layer, for each layer boundary, or for the entire section as a whole in the form of graphs and tables by twelve statistical indicators. The figure below shows a graph of the changes in the average permittivity in a layer of bulk soil. The area of the graph that exceeds the user-defined threshold is colored in red:

Obviously, the transition from the presentation of data on the subsurface environment in the form of a set of amplitudes of the reflected signals (in the form of a radarogram) to the characteristics of this subsurface obtained as a result of applying the BSEF automated analysis method (in the form of an attribute section) significantly increases the informativeness of the GPR study and makes the result of this study more understandable as geoscientist and the customer of the GPR study. Exporting the processing results to graphic formats, spreadsheet formats and a grid data format allows you to work with these results in computer programs to visualize numerical arrays and create drawings, such as Surfer, AutoCAD or similar.

The full list of section attributes in GEORADAR-EXPERT is quite wide - the entire list can be found in the GEORADAR-EXPERT user manual, which is available on Download page. As an example, the figures below show sections along the GPR profile of the attributes of humidity and the Q-factor (the ratio of the center frequency of the signal spectrum to its width). Q-factor is often used to identify loose areas in the ground or areas of reduced strength in building structures.

The indisputable advantage of BSEF automated analysis is the ability to study subsurface environment, the electrophysical characteristics of which change smoothly, without sudden jumps that form reflective boundaries. On the radarogram obtained in the study of such a subsurface, the boundaries of the layers are not visible - as in the figure below to the left. The figure on the right shows a permittivity section along this GPR profile, created in the GEORADAR-EXPERT software:

In the example under consideration, the GPR profile crosses a dried-up stream, the middle of the channel of which is located at a distance of 25 m from the beginning of the profile. The section clearly shows the boundary between dense soil and loose sediments, which, as a result of channel processes, has acquired a funnel-like shape and reaches a depth of 7 meters. An area of high humidity is also clearly visible in the depth range from 2 to 6 meters, which is characterized by increased values of the permittivity and is colored in a dark blue color. The lower part of this area is also funnel-shaped and fairly symmetrical about the middle of the stream. The application of the BSEF automated analysis method to this GPR profile made it possible to obtain much more useful information than the GPR profile could provide before this method was applied.

The runtime spent on performing automated BSEF analysis and creating attribute section is relatively small. For example, the processing of this GPR profile was carried out for several minutes on a computer with a processor frequency of 2.4 MHz. Significant time savings occur in the case of processing a large amount of the same type of GPR data - for example, such as roads or railway profiling results. The user, having configured the parameters of automated processing and starting the processing, can switch to other tasks, and GEORADAR-EXPERT will independently load GPR profiles and save the processing results to the computer`s hard disk.

The GEORADAR-EXPERT software contains a fairly wide range of attributes for various GPR survey purposes. There are frequent cases when information about the object under study is not contained in full by one section of the attribute, but is distributed over the sections of several attributes. In this case, the summation of a set of attributes representing various physical aspects of the object under study allows you to combine disparate information about this object within a single summation section. At the same time, as a result of summation, artifacts caused by the accumulation of errors in the process of collecting and processing GPR information are eliminated. Below are image of the GPR profile obtained during the study of the buried valley, a set of sections of attributes for summation obtained as a result of the automated BSEF analysis of this GPR profile and the result of summing these sections.

The considered GPR profile is not very informative at depths of more than two meters, and the depth of the buried valley, according to a priori information, exceeds this value. On the sections of the attributes, some details of the buried relief are noticeable; however, each of the sections separately does not provide complete information about the structure of the studied subsurface environment. The summation operation made it possible to bring these disparate useful signals into one whole. The relief of the buried valley in the total section is well traced along its entire length. In the figure, the relief of the valley is marked with a dotted line. This example shows how the use of the summation module made it possible to effectively solve the problem of insufficient information content of the GPR profile and individual sections of attributes along this profile.

Using the BSEF automated analysis method has the following advantages over other methods of digital processing of GPR signals:

The depth of GPR research is increasing - the data processing algorithm by the BSEF automated analysis method has high noise immunity and works well in areas of strong noise of signals located in the lower zone of the GPR profile;

The information content of GPR studies increases - a attribute section according to the results of an automated BSEF analysis allows obtaining information on the structure of the investigated subsurface environment even in the conditions of a smooth change in the electrophysical characteristics of this environment, i.e. in the absence of layer boundaries on the radarogram. If these boundaries are present, then a change in the electrophysical characteristics inside each layer also represents undoubted benefit;

Significantly increases the speed of processing GPR data, which is important for constantly increasing volumes of georadar works, especially in the road and railway industries;

The areas of application of the GPR method are expanding and the list of tasks to be solved with the help of this method is increasing;

The influence of the human factor on the processing and interpretation of GPR data is reduced;

Provides greater opportunities for the study of complexly constructed subsurface environments.