Recently Greg Johnston of Sensors and Software gave a very informative webinar on the factors that limit the depth to which ground penetrating radar (GPR) is effective in detecting objects. Greg showed how statistical averaging ("stacking") can be used to increase the signal to noise ratio and substantially increase, even more than double, the depth to which objects can be reliably detected.
The frequency of the radio waves used in GPR is an important determinant of the depth to which GPR is effective. The lower the frequency, the deeper radio waves reach. But there is a trade off, because the lower the frequency the lower the resolution of what is detected. Also the lower the frequency the longer the radar antennas required which can stretch to 8 m at 12.5 MHz to 1 m at 100 MHz to 7 cm at 1 GHz. For example, while a 400 MHz signal may detect objects down to 25 cm, objects less than 6.25 cm in size may not be resolved. Furthermore the deeper the object, the more the radio waves are attenuated, and it may not be possible to resolve objects 5-6 cm in size at depths below 50-60 cm.
GPR can use frequencies in the range 10 MHz to 2 GHz, but typically the frequencies used by commercial products are in the range 100 MHz to 1000 MHz. The webinar focussed on 100 MHz, but the principles determining the depth to which GPR is effective apply to just about any frequency. Because of the low resolution of low frequency GPR, it is not typically used for detecting underground utilities which can often be small diameter cables and pipes lying fairly close to the surface. But for geological investigations, large scale archaeological research or scannng for large underground utilities, low frequency GPR is often the best choice.
According to Greg, the three most important factors determining how deep GPR is effective are
- Electrical conductivity of the soil
- Power of the GPR radio transmitter
- Background RF noise
Electrical conductivity
Radio waves do not penetrate far through soils, rocks and most man-made materials such as concrete. The most important external factor determining the depth achievable by GPR is the electrical conductivity of the subsurface material. For example, while radio waves only penetrate a few millimetres in salt water, they can penetrate 100 metres in granite.
Transmitter power
The FCC in the U.S. and other organizations in other countries limit the power of radio waves emitted by untrawide band devices into the atmosphere, but do not place restrictions on transmissions into the ground. By using shielding to prevent emissions into the atmosphere, the power of the radio transmitter can be increased and greater depths achieved. But this is difficult to achieve at low frequencies and is not a viable way to increase the effective depth.
Radio frequency noise
A major problem when using very sensitive RF antennas is that they pick up noise from other sources such as radio and TV stations, walkie-talkies, and cell phones. Some of these are not random ("coherent external signals"), but most of these are random and comprise background noise. Increasing the ratio of the signal reflected from underground objects to background noise allows weaker signals to be detected which increases the depth to which GPR is effective.
One way to increase the signal to noise ratio is to use statistical averaging. Repeatedly scanning the same area and then averaging the results effectively averages out the background noise and accentuates the signal reflected from underground objects. While one scan may limit the effective depth to 1.8 m, averaging 256 scans can extend the depth to 3 m. This process is referred to as "stacking". Stacking reduces the noise level, increases the signal to noise ratio, and increases the depth of penetration. To take advantage of stacking, the essential technology advance is being able to capture multiple scans rapidly. Sensors and Software's NOGGIN Ultra 100 is able to capture and average between 256 and 65,000 scans. The speed at which the GPR device is moved determines the maximum number of stacks. The DynaQ software implemented on the device automatically calculates the stacking level based on the speed at which the device is moved. A speed of 0.2 km/hr will allow the maximum 65000 scans. For a faster speed the software will limit the number of stacks which reduces the detection depth. Greg gave an example where the effective depth was increased from 5 m to 16 m in sandy soil.
A complicating factor is coherent external signals that are often generated by the GPR device itself. An important technical requirement for a GPR device to be effective at greater depths is that it needs to be "quiet" with respect to extraneous RF emissions.
Statistical averaging has been used in other fields such as seismic exploration for many years. For underground utility detection it has been discussed for years, but until recently technical limitations on the speed of sampling have prevented the development of practical devices and software implementing this technology. Now devices and software implementing stacking technology and available from Sensors and Software and other vendors are enabling more efficient, cost effective, and reliable 3D mapping of underground infrastructure using GPR.
Subscribe
Comments