Data collection through Ground penetrating radars
Let’s talk a bit about how data is being collected when using a GPR and what it means.
As we previously discussed, when doing a GPR survey, the device uses a source antenna (Tx) to send pulses of waves with various frequency into the ground. The waves travel through whatever material is underground and becomes distorted, based on the material’s composite, or dielectric properties.
At certain levels, where the subsurface dielectric properties change suddenly, the waves undergo a transmission, reflection and/or refraction. The receiver antenna (Rx) then measures this distortion of wave signals.
The Rx is sensitive to the electric fields carried by EM wave signals, transforming incoming GPR signals into electrical current. Ultimately, the Rx measures the amplitude and polarization of incoming EM wave signals as a function of time.
Wave signals for GPR travel through the ground at speeds very similar to the speed of light. This means, the travel time for the signal as it propagates from Tx to the ground and back to Rx is extremely short. The time series collected for a single GPR pulse usually takes up to a few hundred nanoseconds after the signal is generated.
Because GPR measurements for a single pulse are so short, they can be repeated multiple times for the same Tx — Rx pair at the same location. This allows the quality the data to be improved by stacking.
Visualization
2D Visualization: Radargram
To gain insights on buried objects, GPR measurements are done at multiple locations using a specific survey config. For common offset, data is typically collected along one or more 2D profiles.
The horizontal axis corresponds to the location of the Rx relative to the Tx for a particular measurement. The vertical axis shows the total travel times of measured signals. By having knowledge of the EM wave velocity, the vertical axis is sometimes represented by an approximate depth. The gray-scale shows the amplitude of the returning signal at each time and location.
3D Visualization
GPR data can be visualized in 3D if it’s collected along multiple profile lines. Below are two examples of this. Each 2D profile is plotted separately. On the bottom right, we see GPR data collected over several buried storage tanks. The image shown is of several horizontal slices.
Processing
Before raw data can be represented as a final image, understandable by humans, several processing steps need to be taken.
Stacking
Stacking describes the process of averaging a set of repeated GPR pulses in order to reduce noise and improve interpretation. Basically, stacking improves the signal to noise ratio for GPR data collected. As we can see from the examples below, the more readings we stack, the clearer we see coherent GPR signals.
Smoothing
Another technique used to improve the interpretation of collected GPR data is smoothing. In GPR systems, the data sampling rate is such that the returning wavelet signals should be reasonably smooth and visible. Random noise on the other hand is completely incoherent. When smoothing is applied to the data, it has the effect of reducing the amplitude of incoherent noise while retaining naturally smooth signals in the data. One example of smoothing techniques is the moving average. We can see that as more data points in time are used for the average, the more easily recognizable the wavelet signal is.
Thanks for reading! Let us know if you have any questions in the comments and be sure to check our website for more info on our products and services.
