Automatic Hiperbola Detection and Apex Extraction Using Convolutional Neural Network on GPR Data
Daffa Dewantara(a*), Wahyudi W. Parnadi (a)

a) Department of Geophysical Engineering, Institut Teknologi Bandung, Jl. Ganesha 10, Bandung, Indonesia.

Abstract

Ground Penetrating Radar (GPR) is a non-destructive geophysical method used for subsurface mapping. It is frequently used for detecting buried cylindrical objects such as underground pipes and cables. Buried cylindrical objects show a hyperbolic signal pattern on a radargram. The typical shape of the hyperbolic reflections depends on the depth and material of the buried objects and the surrounding materials. In many cases, detecting buried cylindrical objects is quite a time-consuming task, thus limiting further interpretation procedures. In this paper, we propose a new method in automating hyperbola detection on radargram, by combining object detection methods through convolutional neural network, and digital image processing techniques. Our work consists of three steps. The first step is pre-processing, which is then followed by converting the data to raster format. In the second step, we used the Faster-RCNN to extract the hyperbola segments as a set of rectangular boundary boxes. The convolutional neural network was trained using synthetic data simulated by the gprMax software. The third step is to estimate the coordinates of the hyperbola apex using a search window algorithm on a digital image. Using these three steps, the detection of buried cylindrical objects using GPR can be automated with a minimal amount of time.

Keywords: Ground penetrating radar- object detection- radargram- convolutional neural network- gprMax

Topic: Earth and Planetary Sciences