13.2.1 Design of the HSR

Subsurface radar designs use classical radar technology principles. The signal emitted into a surveying medium reflects from variations in the electromagnetic properties—for example, heterogeneities, soil layers, buried targets, and so on. The radar system collects the reflected signal in the antenna and amplifies it. After processing, the recorded signal information appears on a computer display.

Figure 13.3 presents two simple schemes for impulse and HSRs. The impulse radar directly collects and amplifies the time-varying reflected signal. Signal processing can estimate the depth of subsurface objects by measuring the time-of-flight of the reflected signal and factoring in the electromagnetic wave speed in the dielectric medium. In contrast, as shown in Figure 13.3, the RASCAN-type holographic radar has a transmitting antenna and two receiving antennas for parallel and cross-polarizations [19]. The RASCAN radar transmits unmodulated, continuous-wave signals. The transmitter generates five discrete frequencies (j_min – j₁, j₂, j₃, j₄, j_max – j₅) covering a bandwidth (j_max − j_min) of about 0.4 GHz. The choice of operating frequency band j_min to j_max depends on the maximum expected target depth in the investigated medium, the medium’s electrical characteristics, and the desired resolution (i.e., target size and/or separation). Three variants of RASCAN-type radar are currently available, with their specifications described in Table 13.1. The need to provide a sufficient contrast in the output image for at least one of the operating frequencies requires the use of multiple frequencies. As Section 13.2.2 will explain, in the case of a single frequency, there can be “blind depths” for a particular target depth due to the sinusoidal variation with depth in the phase difference between the reference and target beams.

TABLE 13.1
RASCAN Holographic Radar Parameters for a Low Attenuating Soil

The RASCAN radar imaging technique works by displaying the analog signal from the mixer output after analog to digital conversion. Therefore, unlike impulse radar, RASCAN has no comparatively expensive radio frequency frontend, and it directly produces plan-view images without the compilation of individual B-scans into a 3-D block model, followed by time-slicing, to produce plan-view images, as is necessary with impulse radar.

Another important distinction between impulse and holographic radar is the type of frequency spectrum. Impulse radar transmits an ultrawideband (UWB) signal, with an essentially continuous frequency spectrum by means of a waveform that typically resembles one and a half periods of a damped sinusoid. On the other hand, HSR has a discrete spectrum, made possible by the continuous-wave signal.

A time-varying amplification in the stroboscopic ISR receiver provides a higher reflected signal amplification for deeper (i.e., longer time-of-flight) objects. The time-varying amplification gives the radar its main advantage of high effective penetration depth. On the contrary, the HSR has the same amplification for objects at all depths since reflections from all depths arrive continuously and simultaneously. Thus the difference in “effective” depth capability of the impulse and holographic systems depends on the mode of operation and the capability of ISR to selectively amplify reflected signals from different depths. The effective penetration depth of HSR systems depends on the attenuation properties of the surveyed medium, the dielectric contrast between the target and medium, and the size and distribution of medium heterogeneities at shallow depths (i.e., above, and possibly concealing, the desired targets).

HSR has an advantage of providing easier compliance with electromagnetic compatibility regulations. In this respect, the time-domain impulse radar has a disadvantage because of the wide, continuous, and difficult to adjust spectrum. The impulse radar spectrum may interfere with other microwave devices, for example, global positioning systems, mobile telephony, electronic switches and fusing, communication systems, and so on. These impulse radar spectrum limitations can conflict with existing electromagnetic compatibility standards.

The USA Federal Communications Commission (FCC) Regulations of 2002 restrict impulse radar spectra as shown in Figure 13.4 [16,17,]. Because discrete frequency holographic radars do not fall under the 2002 restrictions, they can easily be adapted to comply with all regulatory, industrial, or military standards and norms. Furthermore, FCC regulations restrict UWB impulse radar operation to law enforcement, fire and rescue, scientific research institutions, mining companies, and construction companies. A careful reading of the FCC rules (DA-02-1658A1) reveals that they specifically apply to “UWB imaging systems.” Therefore, HSR may not fall under these rules at all.

For shallow scanning, HSR has an additional distinct advantage in plan-view resolution over ISR because large antenna dimensions do not limit spatial resolution. Finally, HSR technology has another important capability—it can image targets in dielectric materials that lie closely above, and even directly on, a metal surface. Such composite materials are very difficult to inspect nondestructively with ISR because of the reverberation of pulses between the impulse radar antenna and the metal substrate. This reverberation obscures the reflections from the target heterogeneities, and renders their actual location and shape indeterminable. When these reverberations occur, multiple reflections (often called ghosts or phantoms) of transmitted impulse signal hide the object of interest [20]. During structural inspections, HSR can image objects on metal surfaces such as concrete flooring poured on corrugated metal pan decking, or moisture between insulating foam and sheet metal structures. This capability can provide a way to inspect spacecraft heat protection shields installed over metal surfaces [14,15]. Table 13.2 compares some characteristics of impulse and HSRs.

The variants of RASCAN radars shown in Table 13.1 can optimize the trade-off between resolution and penetration required for any nondestructive testing application. It needs to make an essential distinction of the term “resolution” when applied to impulse and HSR. For impulse radar, the depth resolution often means the “thin bed resolution” related to the required thickness of a target to separate the reflections from its upper and lower surfaces. Therefore, ISR system designs for resolution must consider the antenna characteristics (gain, divergence, bandwidth, etc.), the received signal aperture size, and the medium dielectric constant. For RASCAN HSR, the term resolution refers exclusively to the planview (X-Y Plane) resolution in the Rayleigh Criterion sense, that is, where the imaging process becomes diffraction-limited when the first diffraction minimum of the image of one source point coincides with the maximum of another. This type the resolution depends primarily on the wavelength in the media.

To record a microwave hologram, the operator moves the RASCAN scanning head line-by-line over the surface of the location under inspection. The special distribution of electromagnetic field of the RASCAN-type radar’s antenna in the near zone produces the high plan-view resolution through holographic processing [21].

TABLE 13.2
Comparison of Impulse and HSR Parameters

13.2.2 Theoretical Analysis for a Point Scatterer

As stated above, RASCAN radars operate at five frequencies essentially simultaneously with respect to the scanning speed. A single frequency can create “blind depths” at some distance where the output of the RASCAN vanishes despite a sufficient dielectric contrast between the object and medium. To demonstrate this, a mathematical model of the simplest monochromatic HSR has to be examined [5].

The response of the RASCAN for the case of a point scatterer buried in a homogeneous medium at a fixed depth h was evaluated. The antenna radiates electromagnetic waves at a constant angular frequency ω whose amplitude and phase do not depend on time. The received wave from the target shown in Figure 13.5 has amplitude A_r (L, ψ), which depends on the range L and the line-of-sight or slant range angle ψ. The phase ϕ_r. depends on the range L to the object as follows:

where ε is the dielectric permittivity of the medium and c is the speed of light in vacuum. In practice, for a noninfinitesimal scatterer, the term A_r depends also on object dimensions, shape, and dielectric properties, according to geometrical theory of diffraction.

By assuming the absence of direct coupling between the transmitting and receiving antennas, we can describe the received signal as a function of time t,

When the received signal of Equation 13.2 combines with the radar reference signal of Equation 13.3 in the mixer, the reference signal with amplitude A₀ and constant phase ϕ₀ results:

The resulting amplitude of baseband signal at the mixer output becomes

From this relation, we can conclude that if the phase shift between the reference and received signal approaches the out of phase cancellation condition

then the level of recorded signal from the object is negligible. This condition produces the blind depth that results from a single frequency signal.

On the other hand, if the phase shift between the reference and received signal approaches the in-phase condition, then the signals add to each other and produce a maximum recorded signal level. This happens when

In RASCAN, the phase term ϕ₀ is different for all five radar frequencies and depends on the antenna design. We have experimentally determined the optimum separation of the five frequencies for each RASCAN model to guarantee a high image object contrast in at least one image for all possible depths, thus lessening the potential for blind depth effects.

To understand the principle of HSR image formation, recall Figure 13.1 that presented an optical analogy to demonstrate the principle of holographic radar image formation. A flat monochromatic reference wave falls on a point object, which scatters it. The result of the interference of the reference and scattered waves on a recording medium plane located at some distance behind the object forms an interference pattern. If the interference plane lies perpendicular to the direction of the reference wave propagation, the interference pattern produces a series of concentric circles like the zones of a Fresnel lens. Papi et al. adopted these concepts in early microwave holographic interferometry experiments [22].

RASCAN HSR radars record the interference pattern along parallel scanning lines and send the spatially sampled data on a USB cable to a personal computer (Figure 13.6). Proper recording of radar holograms with the RASCAN radars requires minimal operator training. Sweeping the radar head produces the holograms as individual raster or scanning lines that are recorded by the data collection system. These lines must be parallel and equidistant to avoid distortion of the image. The time required for the scanning procedure depends on the dimensions of the area, and the selected step between raster lines. Usually for the mid-waveband model RASCAN-4/4000, the step between lines equals 1 cm. Along scanning lines the step between measurement points equals 0.5 cm. This gives typical pixel dimensions of 0.5 cm by 1 cm. Obviously, a smaller step between lines increases the scanning time, but a larger step increases pixel size and decreases the quality of images. The inherent radar resolution at shallow depths (based on signal frequency and antenna design), typically dictates the optimal choice of pixel dimensions. Experiments with the RASCAN-4/4000 experimentally confirmed this for an approximately 2-cm resolution case [21]. The collection of signals along multiple parallel scanning lines forms a raster-based image. The display typically presents the data in gray scale according to Equation 13.4. Typical media such as concrete, bricks, damp soils, and so on have high attenuation for most subsurface radar applications. In these cases, the recorded patterns that form the image appear as a shadowy picture reminiscent of an X-ray image because the media attenuation decreases the intensity of the outer Fresnel rings of the interference pattern. The use of actual scanning examples in Figures 13.7 and 13.8 illustrates and explains this phenomenon.

In the RASCAN-type HSR, the direct transmitter-to-receiver antenna coupling signal does not depend on the scanned medium characteristics and therefore does not influence the demodulation process. To get this advantage, a horn-type antenna was used instead of a bowtie. To form a complete interference pattern, a buried object must reveal its presence to the antenna not only in nadir (ψ = 0) but also in other positions away from nadir, where the distance or range (L) to the object increases. In high-loss media, the signal attenuation along such slant paths away from nadir can be large enough to not reveal any interference pattern. Furthermore, because the RASCAN normally operates in the near field, it has a small aperture compared with the bistatic antenna of impulse-type UWB radar, and is generally geometrically insensitive to contributions from such slant paths. For low attenuation media cases, we see the most complete interference patterns in air, but we do not need this phenomenon for subsurface nondestructive testing radar applications.

Experiments with RASCAN have demonstrated that the main factors that deteriorate the quality of HSR holograms result from attenuation and inhomogeneities in the medium in which the electromagnetic wave propagates. Moisture frequently has the greatest influence on both of these factors. In contrast to water vapor or humidity, which is transparent in the visible spectrum (fog and rain are not included in this consideration), the moisture in soils or construction materials drastically changes their complex permittivity ε [3]. Water itself has an anomalously high real part of permittivity εw′=81 that changes with frequency. The main influence of ε is on the velocity of an electromagnetic wave in the medium—with velocity inversely proportional to ε′. The imaginary part of permittivity ε” is proportional to electrical conductivity of the media. Moisture commonly contains dissolved salts in most soils and construction materials, and this increases their conductivity and attenuation. Concrete and clay represent examples of media that have high levels of conductivity and attenuation in moist conditions.

Although mathematical algorithms for reconstruction of microwave holograms from subsurface radar have been developed, the incompleteness of the interference patterns (due to both attenuation and antenna aperture) will make reconstruction of actual subsurface radar holograms difficult [23]. The task of radar hologram reconstruction requires special consideration of medium attenuation and wavelengths that are comparable with the size of the system. There is no direct analogy between the mathematical models of optical and HSR hologram reconstruction. For recording and reconstruction of optical holograms, we have implied valid assumptions of a clear and homogeneous transmission medium. However, these assumptions do not apply to the HSR case, which makes the reconstruction task very different, and much more difficult (or impossible) for all but the simplest objects (e.g., point targets). However, the incompleteness of interference patterns that prevents full reconstruction has a practical advantage because the HSR interference pattern will retain mostly data from nadir illumination, and thus will strongly resemble the actual subsurface object.

To illustrate the above operational concepts, we can do a mathematical simulation of a point source hologram. The point source reflector makes a simple case demonstrating the concentric ring or Fresnel response of holographic radar. Table 13.3 gives the parameters for this simulation, and Figure 13.7 shows the resulting gray scale image [24]. By observing this image, we can see how the hologram is small and not greater than the antenna aperture. In contrast to the optical hologram for a point source, the HSR antenna aperture and attenuation in the high-loss medium means the radar holograph will have few concentric interference rings. As mentioned above, this phenomenon prevents reconstruction of an ideal point-like scatterer. Instead, reconstruction from the incomplete interference pattern produces a finite size shape with high intensity. Multiple points from an extended target will produce an image approximating the scatterer as illustrated by the following experiments.

13.2.3 Validation of RASCAN Images for a Low-Attenuation Medium

As an illustrative example, a RASCAN-4/4000 was used to record holograms of aluminum letters covered by plasterboard sheets with dimensions of 60 cm by 125 cm, with a thickness of 1.2 cm for each sheet. Dimensions of the word “RASCAN” were 44 cm by 11.5 cm. The aluminum lettering was fixed to a sheet of paper, and placed on a plaster sheet and covered by other plaster sheets one-at-a-time. After addition of each new sheet to the stack, the hidden word was scanned by hand. Every scan included simultaneous recording of 10 radar holograms, at each of five discrete frequencies, and two polarizations per frequency. The scanned area had a dimension of 65 cm by 28 cm.

TABLE 13.3
Point Source Holographic Simulation Parameters

Figure 13.8 shows the eight holograms taken with from one to eight 1.2-cm plaster sheets. From the entire array of all received images, only the holograms recorded at a frequency of 4 GHz in parallel polarization have been presented. In the first three images, which were recorded through 1, 2, and 3 plaster sheets respectively, the word RASCAN is legible. If the number of sheets is more than three (depth > 3.6 cm), the outlines of letters become blurred, and the images display a classical holographic interference pattern with wavy nature.

It is possible to explain these phenomena as follows. At very shallow depths, there is direct reflection in nadir from surface of the letters with very high level of reflected signal (higher than the level of the reference signal and higher than off-nadir reflections). With more sheets over the word, the radar antenna records reflections from letters on inclined angles (off-nadir) at a signal level that is comparable with the reference signal. This case clearly shows the wavy nature of the radar images. The RASCAN image recorded through eight sheets (a depth of about 10 cm) resembles the optical hologram that was presented by D. Gabor in his classic work [4]. Figure 13.9c shows Gabor’s optical hologram.

Examining Figures 13.8 and 13.9 gives a clear qualitative comparison between optical and radar holograms. The main distinction between optical and radar holograms is the number of Fresnel zones that may be visible. Typical optical holograms may have 10⁴-10⁵ diffraction lines [6]. HSR holograms have only a few lines. As mentioned above, this is explained by the fact that the system dimension to wavelength relation (d/λ) is much larger for an optical hologram than for an HSR one. However, the absence of many Fresnel lines causes the microwave holographic interference pattern to resemble the target—making hologram reconstruction both difficult and (fortuitously) unnecessary.

Figure 13.10 shows the effect of signal attenuation along the longer (off-nadir) reflections with concomitant loss of Fresnel lines. In this figure, the signal level A was recorded along a single scan line, perpendicular to a hidden wire, for three cases:

1. The wire placed in the middle of a stack of eight dry plaster sheets at a depth of 5 cm

2. The wire placed between two dry foam concrete bricks also at a depth of 5 cm

3. The wire placed between two damp foam bricks, again at a depth of 5 cm

This figure clearly demonstrates the reduction of side lobes or Fresnel lines as the attenuation of the electromagnetic wave in the medium increases. In numbers, the ratio of the main peak height to side lobes for the three cases is approximately 14, 4.3, and 2.8, respectively.

Turning to more practical (rather than simply illustrative) examples, a RASCAN-4/4000 produced the microwave holograms in Figure 13.11. These are images recorded over a stack of seven dry plaster sheets (e_r = 3, σ = 0.003 S/m) with an aggregate thickness of 10.5 cm, where we placed several metal coins of 25-mm diameter, a 3 by 3 cm empty space, and two thin metal wires at different depths inside a 0.5 × 0.5 m area [19]. These images show frames from an animation in which the holograms at different frequencies transform smoothly into one and other. You can view the animation at http://www.rascan.com/RASCAN_test_bed.avi, which illustrates the interference patterns as concentric rings visible around the coin and void targets. The void is the target furthest inside the acute angle formed by the wires. One coin, in the top-left corner of the image, was placed at some distance below the wire, and this is visible in several of the images. Note also that at certain frequencies, the coin shows clearly where the wire above it is barely visible. This shows the strong advantage for holographic radar because the wire would shadow the coin on impulse radar images. Because of the sinusoidal change in phase difference between the reference and reflected signal for objects at different depths, the HSR can achieve higher contrast for one depth or another by cycling through the frequencies in the selected bandwidth. A RASCAN operator typically uses this sort of animation to evaluate the scanning results in near real time.

13.3.1 RASCAN in Building Structure Surveying

The RASCAN was used in the historical former senate building under renovation for use by the Constitutional Court of the Russian Federation in Saint Petersburg, Russia. The outstanding architect K.I. Rossi in 1829-1834 designed this building, and it has great value in Russian culture. As part of an earlier renovation, the engineers installed a radiant heating system in the floor, with the heating pipes covered by cement. The ambiguous as-built heating system specifications presented a significant threat of accidental damage when laying a proposed parquet floor. The contractors needed to determine the exact location of the heating pipes before floor restoration could start.

The heating system reportedly followed this construction sequence: first, the workers covered the subfloor with an elevated metal mesh with cells typically 15 × 15 cm square. They then secured the heating pipes to the mesh by plastic clips. Various heating pipe materials included cross-linked polyethylene (PEX), multilayer (a composite of PEX, aluminum, and PEX) and polybutylene (PB). It was known the pipes had 30-cm spacing between them. A finishing layer of cement typically extending 3 cm above encased the pipes and mesh.

Because the plastic piping has a lower permittivity contrast with the cement than does the metal mesh, we feared that the plastic pipes of the heater system would be invisible on the background of the metal mesh. As explained in Section 13.2, the object contrast in RASCAN images depends on the reflectivity of the object as well as the phase shift between the reference and reflected beams (which depends on the distance to the object). As shown in this example, for the case of extended or elongated objects, the orientation of the polarized signal reflected from an object relative to the receiver also has great influence on the image.

The floor subsurface was inspected by scanning a total surface area of 16.7 m² with the RASCAN-4/2000 holographic radar shown in Figure 13.12. The work (disregarding the time for equipment deployment) took about 5 h. More than half of that time was spent on scanning whereas the rest was spent on plotting the layout of pipes and cables directly on the floor based on the scanned images. For this type of application, a postprocessing approach for image enhancement has been proposed in [25]. While inspecting the floor, we also detected a subsurface tangle of power and communication cables, which added to the complexity of interpreting the radar images.

The surveyed area was divided into 1.7 m by 2.0 m sections. After recording a radar image of each section, the operator analyzed the image and drew the results on the floor with chalk. Figure 13.13 shows the position of heater pipes marked by blue chalk; red chalk was used for marking cables and wires. Examination of the bottom image of Figure 13.14 reveals how parallel polarization failed to detect the horizontal components of the metal mesh and the heater pipes. However, in the radar image, cross-polarization clearly showed the plastic pipes (horizontal) against the vertical lines of the metal mesh. Notice that the cross-polarization case has lower contrast than in the parallel polarization, but the high accuracy of phase variations allows us to correctly interpret the layout of the heater pipes.

We demonstrated another example of how the HSR can aid in historical building preservations by taking RASCAN to Franklin and Marshall College in Lancaster, PA, USA. This is one of the oldest colleges in the United States, and was founded using a grant from Benjamin Franklin. The concrete floor over a corrugated metal pan presented a new challenge. We worked in one of the oldest buildings, Fackenthal Hall during a recent renovation. A floor restoration in the 1930s used corrugated metal pan decking covered with about 10 cm of concrete as shown in Figure 13.15. Note that in this situation, although we could access both sides of the floor, the metal pan deck inhibits use of X-ray techniques for imaging subfloor elements.

An initial scan was performed over an area of 0.5 m by 2 m from the upper surface of the floor. Figure 13.16 shows a single frequency (2.0 GHz) of the RASCAN-4/2000 images taken at two polarizations. In this case, we recorded the scan lines perpendicular to the corrugations. An unexpected result came when the RASCAN radar detected the electrical conduit oriented at an angle to the rebars. An earlier test using a 60-Hz live power line detector (Radiodetection, Ltd. “CAT”) showed no response to this conduit due to it being either dead or shielded by a grounded metal housing. The cross-polarization image clearly showed the four rebars within the scan area.

Another investigation on a smaller area of the same floor, which had two 16-cm diameter holes filled with concrete, has been done. In the past, stove or water pipes had passed through these two holes. During an earlier renovation, workers had removed the pipes and filled the holes with concrete. Figure 13.17 shows RASCAN-4/2000 images, which clearly identify the locations of the two filled holes. The different contrast results either from electromagnetic property variation between the patching concrete and the original flooring concrete or from differing thicknesses of the two, which made them visible to the RASCAN system. The cross-polarization images also showed the buried corrugations in the decking as identified by their correct spatial separation of about 13 cm.

13.3.2 Detection of Water Intrusion in the Underground Parts of a Building

Because of the abnormally high dielectric permittivity of water (ε_r = 81 as opposed to less than 10 for most construction and earth materials), areas with increased moisture content have a high contrast on radar images. This was confirmed during experiments using RASCAN radar to detect locations of groundwater intrusion into the structure of an underground garage as shown in Figure 13.18a.

The microwave image presented in Figure 13.18b shows two inclined, bright, elongated zones that we interpreted as voids through which water advanced vertically downward. To validate this interpretation, a hole was drilled at one of the spots and water immediately flowed from the hole.

13.3.3 Security Applications

RASCAN radar can detect listening or bugging devices intended for clandestine recording of confidential information in government, business, or residential buildings. Owing to the high resolution of RASCAN radar, it can find suspicious areas in different parts of a building and even classify different types of bugging devices from the recorded microwave images. To demonstrate this capability, the model wall shown in Figure 13.19 made of chipboard panels with hidden bugging devices: two microphones and a tiny TV camera was built. Figure 13.19a shows the model, which was covered by two layers of chipboard with total thickness of 5 cm, from the top. Figure 13.19b shows a RASCAN-4/4000 image of the model. The recorded microwave image precisely represents the shape of hidden objects beneath the chipboard plates. This allows discrimination of bugging devices from innocuous objects that may normally occur in building designs, such as nails, reinforcements, and so on.

RASCAN radar could provide a safe and effective tool to complement, for example, nonlinear junction detectors in searches for bugging devices. A RASCAN radar scan on suspicious places found by nonlinear junction detectors would significantly lower the level of false alarms produced by the latter when used alone. RASCAN surveys can also provide an extra check for hidden devices of new and unknown type. This application is ideal for the RASCAN-4/7000 operating in the frequency range of 6.4-6.8 GHz, which has the highest resolution (see Table 13.1). Experiments conducted with this device have shown sufficient sensitivity to detect fiber-optic cables without any metal content.

13.3.4 Humanitarian Demining

We have also experimented with RASCAN as a sensor for humanitarian demining operations and shown a capability to detect metallic and nonmetal mines [26–28]. Figure 13.20 shows the outdoor test bed at the University of Florence where we scanned for metal and nonmetal mine simulators. Figure 13.21 shows the resulting radar images. For comparison, we scanned the same simulated mine field area with a RASCAN-4/4000 HSR and a GSSI 1.5 GHz ISR. Figure 13.21a shows the time-slice (depth-slice) image produced by the GSSI system at a depth between 2 and 5 cm. Note that because the GSSI system uses UWB impulse radiation, it requires considerable postprocessing to first migrate the characteristically hyperbolic reflections on each scan line or B-scan to near-point targets, then combine the bidirectional B-scan profiles into a 3-D data block, and then extract the plan-view depth slice. In contrast, Figure 13.21b shows the raw (no processing necessary) RASCAN-4/4000 image at 3.9 GHz with parallel polarization generated in real time as the scan lines were swept. This shows the type of real-time image available to a field deminer. However, postprocessing can be applied to further improve HSR images. Figure 13.21c shows a composite RASCAN image produced by removing a best-fitting planar background from each of the five individual discrete frequency images and then calculating the root-mean-square (RMS) RASCAN response pixel-by-pixel for all the five. Applying a 3 × 3 Gaussian smoothing produces the final image, which clearly shows the mine locations.

In this test, although both HSR and ISR detected the mine simulators and the wire, the RASCAN images, particularly the data fusion image in Figure 13.21c, clearly reproduce the mine shapes and dimensions. Note that the PMA-2 simulator in the bottom-right corner of the test bed unfortunately fell just outside of the scanned area for all images. In addition, an experimenter’s knee print appears on the right edge of the RASCAN images. This shows the tremendous HSR sensitivity to soil conditions—in this case compaction.