Inspection of changing processing parameters in GPR data interpretation

Ground Penetrating Radar (GPR) is one of high resolution geophysical methods for locating missed objects, buried equipments. GPR is a nondestructive testing method that is useful in various kinds of explorations such as mapping the archaeological sites. The safety of the archaeological sites is one of the most concerns of the archaeologists. One of the best GPR applications is in archaeology as a non-destructive exploration method.
This method uses high frequency electromagnetic waves or Radio Frequency range (12.5–2300 MHz) to investigate the subsurface and detecting underground objects. It is used to measure features such as depth of the buried anomalies, layers thickness, moisture content, horizontal cracks, voids and hidden objects. The electromagnetic waves reflected from the interfaces in the subsurface are used to analyze the structure of the area under scan. The electromagnetic waves are emitted from an antenna that is not necessarily in direct contact with the earth surface. The same antenna can be used to record the reflected energy stemming from the dielectrical interfaces. The penetration depth of the GPR technique varies with acquisition and subsurface conditions.
The data gathering procedure can be designed in two or three dimensions. If the target is such as pipe or the pre-known target, the data gathering can be done in 2D. But if the target is completely unknown, especially as the matter of its place, the three dimensional data grid will be more useful and accurate. The GPR signals diffract when they contact the anomalies. The anomalies seem as hyperboles in sections. With these hyperboles in the sections, accurate detection of anomaly's depth and other properties are impossible. For solving this problem, a processing method should be applied.
The processing methods vary depends on the different goals. Using various processing methods with parameters can be applied for different purposes. The operator's ability in choosing the proportionate parameters is as important as choosing an appropriate processing algorithm to aim for an acceptable image. Although in some special conditions, some processing may cause unwanted problems in data.
This paper contains investigations in an archaeological site in Kerman Province, Iran. The exact location of the site is in the middle of the Kerman city, the capital of the Province. The antenna which has been used was a 500 MHz GPR antenna, made by MALA.
The target was the remains of one the most important and also old schools of Iran which had been buried under the soil for hundreds of years. On the top of the data collection areas, where had been specified by the archeologists, some houses and buildings had been made. The data gathering procedure in the mentioned area was a three dimensional one. The data were collected in 2009. The Reflex 3D software was used for the processing and the depth estimation of the acquisitioned data.
Regarding the properties of the target in the study area, the processing algorithms can be different from case to case. For example, the data filtering that is usually being used for windowing the environmental noises in some cases, is sufficient for bolding the subsurface reflections. Although in some other cases, more and complicated levels of processing are needed.
The whole processing algorithm depends on the land’s properties, the GPR instruments, the softwares and mainly the data gathering methods. So, in all of the cases not all of the processing levels are necessarily needed. Perhaps in some situations, some of the processing levels are not useful and may cause the operator unwanted difficulties in the data sections. In this paper, the needed processing levels in archaeological cases were studied, and also the results on the Kerman area have been presented.
The gain filter was the most sufficient processing step in the processing algorithms. The changing of different parameters in this filter caused the most clear improvement in the results and for making the anomaly’s source known. So, in the cases that the depth range of anomaly’s source had been guessed by the experts or the other methods and the archaeological areas, a proper gain filter was recommended.
One of the other methods that will be so useful in making the anomalies more clear in the GPR sections is the amplitude limitation. The results of applying this method are shown in this paper. The unwanted amplitudes are trouble-making in the GPR anomaly interpretations. The amplitude limitation helps the operator and the interpreter to distinguish the main anomalies from the environmental noises.
The Dewow processing step is also one of the steps which will decrease the near surface or air signals between the transmitter and the receiver. As conclusion, two mentioned processing methods above, namely the gain filter and amplitude limitation, were used as two major parts of processing algorithms and helped the exploration team to clarify the source’s exact coordination.