GPR (Ground Penetrating Radar) is well-known as an effective non-invasive
imaging approach for shallow nature underground discovery, like finding and locating
submerged objects. Although GPR has achieved some success, it is difficult to
automatically process GPR images because human experts must interpret GPR images
of buried objects. This can happen due to the possibility of a variety of mediums or
underground noises from the environment, especially rocks and roots of trees. Thus,
detecting hyperbolic echo characteristics is critical. As a result, Viola Jones detection is
used to determine whether the presence of a hyperbolic signature underground indicates
a pipe or not. GPR can also be used in the public works department because it is a non
destructive tool. Workers, for example, should be aware of the pipe size that must be
replaced when it leaks. The original GPR image already shows hyperbolic image
distortion due to pipe refraction. The current method is unreliable due to its lack of
flexibility. As a result, there is another method for resolving this issue. Thus, the image
will be pre-processed to eliminate or reduce background noise in the GPR input image.
The results of this project demonstrate that the Viola Jones algorithm can accurately
detect hyperbolic patterns in GPR images.
