GPR Mapping in Excavation Strategies

Why GPR Data needs to be processed:

Typical GPR reflection profiles contain noise and unwanted reflections that are produced by:
- antenna "ringing"
- differences in the coupling of energy with the ground
- multiple reflections that occur between the antenna and the ground surface
- system and background "noise"

Background Removal is the first processing stage, which will remove the horizontal lines common in profiles.

Horizontal lines in reflection profiles are caused by much of the "noise" listed above. They obscure the "real" reflections and must be removed.
Figure 28
The removal process is a simple arithmetic calculation that can be done with digital data. All reflections that occur at the same time along the whole profile are averaged, and this average is subtracted from all the individual reflections that make up the profile. When this is done the profile is "cleaned up" and contains only the non-horizontal reflections, which are those that come from within the ground.
Figure 29

Point Source Hyperbola Tail Removal

Reflection profiles also contain high-angle reflections that are the tails of hyperbolas generated from buried rocks and walls. Although these reflections can be helpful in visual interpretation, because they show the tops of walls and individual rocks, they tend to make profiles "busy" and produce blurry maps when processed into amplitude slices.
The tails of the hyperbolas can be removed, leaving only the reflections at the apex, which represent the true location where the top of the feature is buried.
Figure 30
To illustrate the difference between maps that have been processed to remove the hyperbola tails, and those that are not, a comparison was done on slices in Grid 2. In Figure 31 maps are compared that were constructed with the tails still in the data, and after they were removed. Those with the tails removed are much more accurate, as they do not contain the spurious reflections of the hyperbola axes.
Figure31

Production of Rendered Images of Processed Reflections

Once data are processed by removing unwanted noise, and the point source hyperbola tails are removed, the resulting reflection profiles are ready for additional image production.
Thirty-five horizontal slices were made of the data in Grid 2. These data were then imported into a volumetric-data visualization program that can isolate amplitude anomalies and render them into three-dimensional shapes. When this was done for all the highest amplitudes in Grid 2 a three-dimensional image of buried northern structure was produced, which was then shaded to enhance the reflections. This image mimics the actual location of the structure in the ground. To play this video, click on the thumbnail image below:
Figure 32