**2.1. Study area: Mustafa Kemal ATATURK'S Mausoleum, ANITKABIR**

The Anitkabir mausoleum, was completed in 1953 in Ankara, Turkey (Figure 1). It hosts state ceremonies during national festivals, and represents the Turkish people and Ghazi Mustafa Kemal ATATÜRK, the founder of the Republic of Turkey (Figure 1b). Anitkabir was con‐ structed in three phases. The first part is an entrance road, called the Lion Road, which is 262m long and has a total of 24 lion statues along each side, representing power and peace (Figure 1c). The second part is a ceremonial square, and the third is the Mausoleum (Figure 1c). At the beginning of the Lion Road, the Turkish people are represented by three large male statues in front of the Freedom Tower on the left side (Figure 1d), and three large female statues in front of the Independence Tower (Figure 1e) on the right side [29].

The monument groups (three women, three men) and twenty-four lion statues of Anitkabir are mainly composed of white travertine from Pinarbasi, Kayseri, Turkey. The white-coloured travertine has a banded and spongy texture under the microscope [23]. It is mainly composed of calcite, aragonite with a small amount of salt, recrystallized calcite, gypsum and plant relicts. Table 1 shows the modal mineralogical composition and physical properties of this travertine. Micro-fractures were observed under a polarizing microscope, especially at the rim of the vesicular of the rocks (Figure 2).


**Table 1.** Mineralogical composition and physical properties of white-coloured travertine in Anitkabir.

**Figure 1. (a)** Geographical location of the study area, **(b)** Ghazi Mustafa Kemal ATATURK, the founder of the Republic of Turkey, **(c)** the ANITKABIR monument, Ankara-Turkey, **(d)** three large male statues at the beginning of the Lion Road, which is the entrance to Anitkabir, on the left side, **(e)** female statues facing the male group, **(f)** 24 lion statues, representing power and peace, sit on each side of the Lion Road.

**Figure 2. (a)** A lion statue, **(b)** view close to surface of lion, and **(c)** microphotograph of white-coloured travertine.

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage

http://dx.doi.org/10.5772/54998

111

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage http://dx.doi.org/10.5772/54998 111

**Figure 2. (a)** A lion statue, **(b)** view close to surface of lion, and **(c)** microphotograph of white-coloured travertine.

**Figure 1. (a)** Geographical location of the study area, **(b)** Ghazi Mustafa Kemal ATATURK, the founder of the Republic of Turkey, **(c)** the ANITKABIR monument, Ankara-Turkey, **(d)** three large male statues at the beginning of the Lion Road, which is the entrance to Anitkabir, on the left side, **(e)** female statues facing the male group, **(f)** 24 lion statues,

110 Imaging and Radioanalytical Techniques in Interdisciplinary Research - Fundamentals and Cutting Edge Applications

representing power and peace, sit on each side of the Lion Road.

**Figure 3.** Data measurements on the female statues.

#### **2.2. GPR survey description**

The human statues were divided into several subparts such as skirt or under waist, between waist and neck, arms, trousers, etc. for GPR survey. Some parts were divided into additional subparts (Figures 3 and 4) according to their figures, in order to enable true 3D imaging and protect the profile line position, because topography correction was not possible on the statues. In addition, three profiles spaced 10 cm apart were arranged along the backs of lion statues. The data acquisition scheme is shown in Figure 5, in which GPR data profiles gathered on the body of the first female statue were split into two parts, called skirt and upper part between waist and neck.

Profiles were spaced at 10 cm on each subpart, and were lined with a paper band sticker. First, data survey tests were carried out to determine the recording time-window according to the approximate thicknesses of the statues. A RAMAC CU II GPR system equipped with a 1.6- GHz bistatic shielded antenna was employed on all the statue groups. Transmitter–receiver antenna offset was 0.05m. Trace spacing was 0.0044m and time-sampling interval per trace on each profile was 0.0327 ns.

**Figure 4.** Data acquisitions on the male statues.

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage

http://dx.doi.org/10.5772/54998

113

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage http://dx.doi.org/10.5772/54998 113

**Figure 4.** Data acquisitions on the male statues.

**Figure 3.** Data measurements on the female statues.

The human statues were divided into several subparts such as skirt or under waist, between waist and neck, arms, trousers, etc. for GPR survey. Some parts were divided into additional subparts (Figures 3 and 4) according to their figures, in order to enable true 3D imaging and protect the profile line position, because topography correction was not possible on the statues. In addition, three profiles spaced 10 cm apart were arranged along the backs of lion statues. The data acquisition scheme is shown in Figure 5, in which GPR data profiles gathered on the body of the first female statue were split into two parts, called skirt and upper part between

112 Imaging and Radioanalytical Techniques in Interdisciplinary Research - Fundamentals and Cutting Edge Applications

Profiles were spaced at 10 cm on each subpart, and were lined with a paper band sticker. First, data survey tests were carried out to determine the recording time-window according to the approximate thicknesses of the statues. A RAMAC CU II GPR system equipped with a 1.6- GHz bistatic shielded antenna was employed on all the statue groups. Transmitter–receiver antenna offset was 0.05m. Trace spacing was 0.0044m and time-sampling interval per trace on

**2.2. GPR survey description**

waist and neck.

each profile was 0.0327 ns.

#### **2.3. GPR data processing**

Data processing was performed on 2D GPR profile data sets for each part of the statues with the REFLEXW program (ver. 5.5) developed by Sandmeier Scientific Software [30]. Start-time correction, then de-wow and background removals were applied to all the profile data in order to protect the true time scale, remove very low frequency effect and average amplitude knowledge respectively. The amplitude decay compensation was applied to all traces of the whole data set by using the same small-scale linear gain function. A second-order band-pass Butterworth filter was used for the whole data set to eliminate low-frequency artifacts and high-frequency noise. The resulting synthetic hyperbolas were matched with diffraction patterns throughout the profiles to determine average velocity of the electromagnetic (EM) wave. The best matching hyperbola provided the velocity of 0.12 m/ns. Finally, Kirchhoff migration was applied to the radargrams using average velocity, in order to carry diffracted electromagnetic (EM) waves true locations.

The quality of GPR images is strongly dependent on appropriate correction of the attenuation effects, usually supplied by time-varying gain. However, historically, the use of amplitude gain in basic processing of GPR data has been highly subjective and also very much displaying methodology [31]. There are various methods available for amplitude gain for GPR data. Traditional time-varying gain is carried out using linear, exponential functions, etc. functions, including ground wave amplitudes. However, this operation is not linear. The time-gained GPR data cannot recover the original information. Selection of the gain function depends mostly on the user and the quality of the GPR data. Both exaggerated linear and exponential time-gain change not only the amplitude range for each time step but also amplitude shape. Exaggerated time gain to image 2D GPR data can result in erroneous interpretation, especially when using such data to construct a 3D volume.

Time-gained signal is

$$s(\mathbf{x}, t) = r(\mathbf{x}, t)h(t) \tag{1}$$

$$h(t) = a, 1 < a \le \left| \frac{r\_{\text{max}}}{r\_{\text{min}}} \right| \tag{2}$$

**Figure 5.** GPR data acquisition scheme of the first female statue.

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage

http://dx.doi.org/10.5772/54998

115

Where, *h* (*t*) , is a time gain function, |*r*max | and |*r*min | are maximum- and minimum amplitude of the reflected/diffracted wave *r*(*x*, *t*), respectively.*h* (*t*), has a constant decimal value *a* between 1 and 2, and also is a linear function between 1 and *a* for a 2D data profile [32].

Our approximation concerns 2D or 3D image simplification in determining buried archaeo‐ logical remains and discontinuities such as fractures. Therefore, we assigned a new amplitude– colour range for 2D radargrams of the GPR profiles and opaque range in order to identify anomalies and display the data set in a transparent 3D data volume. Figure 6a indicates some processed radargrams of the profiles of the skirt of the first female statue, shown in Figure 5. **2.3. GPR data processing**

electromagnetic (EM) waves true locations.

when using such data to construct a 3D volume.

Time-gained signal is

Data processing was performed on 2D GPR profile data sets for each part of the statues with the REFLEXW program (ver. 5.5) developed by Sandmeier Scientific Software [30]. Start-time correction, then de-wow and background removals were applied to all the profile data in order to protect the true time scale, remove very low frequency effect and average amplitude knowledge respectively. The amplitude decay compensation was applied to all traces of the whole data set by using the same small-scale linear gain function. A second-order band-pass Butterworth filter was used for the whole data set to eliminate low-frequency artifacts and high-frequency noise. The resulting synthetic hyperbolas were matched with diffraction patterns throughout the profiles to determine average velocity of the electromagnetic (EM) wave. The best matching hyperbola provided the velocity of 0.12 m/ns. Finally, Kirchhoff migration was applied to the radargrams using average velocity, in order to carry diffracted

114 Imaging and Radioanalytical Techniques in Interdisciplinary Research - Fundamentals and Cutting Edge Applications

The quality of GPR images is strongly dependent on appropriate correction of the attenuation effects, usually supplied by time-varying gain. However, historically, the use of amplitude gain in basic processing of GPR data has been highly subjective and also very much displaying methodology [31]. There are various methods available for amplitude gain for GPR data. Traditional time-varying gain is carried out using linear, exponential functions, etc. functions, including ground wave amplitudes. However, this operation is not linear. The time-gained GPR data cannot recover the original information. Selection of the gain function depends mostly on the user and the quality of the GPR data. Both exaggerated linear and exponential time-gain change not only the amplitude range for each time step but also amplitude shape. Exaggerated time gain to image 2D GPR data can result in erroneous interpretation, especially

( ) max

,1 *<sup>r</sup> ht a a*

min

*r*

Where, *h* (*t*) , is a time gain function, |*r*max | and |*r*min | are maximum- and minimum amplitude of the reflected/diffracted wave *r*(*x*, *t*), respectively.*h* (*t*), has a constant decimal value *a* between 1 and 2, and also is a linear function between 1 and *a* for a 2D data profile [32].

Our approximation concerns 2D or 3D image simplification in determining buried archaeo‐ logical remains and discontinuities such as fractures. Therefore, we assigned a new amplitude– colour range for 2D radargrams of the GPR profiles and opaque range in order to identify anomalies and display the data set in a transparent 3D data volume. Figure 6a indicates some processed radargrams of the profiles of the skirt of the first female statue, shown in Figure 5.

*s xt r xt ht* ( , , ) = ( ) ( ) (1)

= <£ (2)

**Figure 5.** GPR data acquisition scheme of the first female statue.

A linear colour scale of the radargrams (Figure 6b) indicates the amplitude range, which begins with maximum negative polarity and ends with maximum positive polarity. The blue colour range represents maximum negative amplitude, while the purple colour range represents the positive amplitudes according to their values. It is known that the maximum positive and maximum negative amplitude ranges in the amplitude–colour scale represent the fractures and cavities filled with air or buried archaeological remains in soil. Therefore, it is necessary to check the maximum amplitude ranges on the time slices to identify these target objects. We applied a new approximation to eliminate the weak reflections that are characteristic of cemented rock [7, 23], and to activate the fractures and cavities represented by strong reflec‐ tions/diffractions on the radargrams (Figure 6c). This involved assigning a new colour scale for the amplitude range of the processed profile data set by means of a new amplitude–colour

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage

http://dx.doi.org/10.5772/54998

117

**Figure 7. (a)** A linear functioned amplitude–colour scale and **(b)** re-functioned colour scale of the same amplitude range.

**Figure 8. (a)** A linear functioned opacity for the linear functioned amplitude–colour scale and **(b)** A re-arranged opac‐ ity function using the same amplitude–colour scale to activate only maximum amplitude range and remove all others.

function rather than linear amplitude–colour function (Figure 6d, Figure 7).

**Figure 6.** a) Processed 2D radargrams of the GPR profiles 1 to 5, acquired from the skirt of the first female statue, using (b) selected linear amplitude–colour function, (c) the same radargrams using (d) re-arranged amplitude–colour scale of (b) to reveal small cavities and fractures on the radargrams.

A linear colour scale of the radargrams (Figure 6b) indicates the amplitude range, which begins with maximum negative polarity and ends with maximum positive polarity. The blue colour range represents maximum negative amplitude, while the purple colour range represents the positive amplitudes according to their values. It is known that the maximum positive and maximum negative amplitude ranges in the amplitude–colour scale represent the fractures and cavities filled with air or buried archaeological remains in soil. Therefore, it is necessary to check the maximum amplitude ranges on the time slices to identify these target objects. We applied a new approximation to eliminate the weak reflections that are characteristic of cemented rock [7, 23], and to activate the fractures and cavities represented by strong reflec‐ tions/diffractions on the radargrams (Figure 6c). This involved assigning a new colour scale for the amplitude range of the processed profile data set by means of a new amplitude–colour function rather than linear amplitude–colour function (Figure 6d, Figure 7).

**Figure 7. (a)** A linear functioned amplitude–colour scale and **(b)** re-functioned colour scale of the same amplitude range.

**Figure 8. (a)** A linear functioned opacity for the linear functioned amplitude–colour scale and **(b)** A re-arranged opac‐ ity function using the same amplitude–colour scale to activate only maximum amplitude range and remove all others.

**Figure 6.** a) Processed 2D radargrams of the GPR profiles 1 to 5, acquired from the skirt of the first female statue, using (b) selected linear amplitude–colour function, (c) the same radargrams using (d) re-arranged amplitude–colour

116 Imaging and Radioanalytical Techniques in Interdisciplinary Research - Fundamentals and Cutting Edge Applications

scale of (b) to reveal small cavities and fractures on the radargrams.

limitation allowed simplification and made only fractures and native cavities visible on the radargrams, as seen in Figure 6c. The horizontal x-axis of Figures 6a and 6c indicate the distance along the measuring profile. The vertical axis shows depth range, which represents the thickness of the statue skirt from front surface to back surface, transformed by using average

Transparent 2d/3d Half Bird's-Eye View of Ground Penetrating Radar Data Set in Archaeology and Cultural Heritage

http://dx.doi.org/10.5772/54998

119

Generally, interactive visualization is carried out by constructing 3D data volumes of parallelaligned 2D GPR data sets to show the target objects. The 3D data volume can be displayed as slices, including profiles, times (or depths) and common traces of the profiles; or separated sub-blocks are rendered as solid iso-volumes with linear opacity, determined by the amplitude of the anomalies. The buried fractures or cavities can be defined on the interactive slices, particularly on depth slices with location, and shapes according to depth. Therefore, it was necessary to check the most meaningful depth slices and profiles to define the structures of the subsurface if the area is small and complex. However, the obtained results could be further

Our aim was to obtain a good 3D data volume display, which was a critical part of interpreting the GPR data set. The 3D image is able to present a view of subsurface features such as a fracture or cavity, in addition to objects such as industrial and/or archaeological remains, etc. This imaging could be achieved by a transparent 3D half bird's-eye view revealing only buried objects. Therefore, firstly, transparency could be achieved by constructing an opacity function instead of linear opacity determined by the amplitude scale (Figure 8). The horizontal axis of the opacity function was the amplitude scale starting with maximum negative amplitude and ending with maximum positive amplitude; the vertical axis represented opacity coefficients of the amplitude range [11, 23]. Thus, any amplitude range could be highlighted or minimized by the appointed opacity coefficient. The REFLEXW program allows the opacity coefficient to be chosen between one (maximum opacity) and zero (transparent) (Figure 8b) [26]. A trans‐

parent view could be obtained only by eliminating the unwanted amplitude range.

location and shapes according to depth (thickness).

Therefore, the amplitude range was important. Because it was known that the maximum amplitudes represented discontinuities, the weak amplitude range was eliminated by giving these a zero opacity value, and transparent 3D imaging was obtained. The transparency was achieved by allocating an opaque interval to the amplitude scale, similar to the re-arranged amplitude–colour approximation for interested profile range or time range for the solid 3D GPR data volume. This visualization type was applied to both the statues and archaeological remains in this chapter. Figure 9a indicates traditional, solid depth-slices, while Figure 9b indicates transparent depth slices at 15 cm, 38 cm and 51 cm of the data set for the skirt of the first female statue. These slices were used to control micro-fractures and cavities according to the skirt thickness. The horizontal x-axis and y-axis of slices indicate the profile sequence and the distance along the measuring profile respectively. Locations of the micro-cavities could be seen on the slices. It is necessary to carefully check interactive slices in order to determine

velocity of the EM wave.

improved.

**2.4. Transparent 3D half bird's-eye view of GPR data set**

**Figure 9. (a)** Traditional solid depth slices at 15 cm, 38 cm and 51 cm, **(b)** Transparent depth slices for the skirt of the first female statue (same three depths).

The horizontal axis of the amplitude–colour function (Figure 7) is the amplitude scale of the GPR data, whereas the vertical axis represents colour categories from 0 to 255. The colour limitation allowed simplification and made only fractures and native cavities visible on the radargrams, as seen in Figure 6c. The horizontal x-axis of Figures 6a and 6c indicate the distance along the measuring profile. The vertical axis shows depth range, which represents the thickness of the statue skirt from front surface to back surface, transformed by using average velocity of the EM wave.
