Please use this identifier to cite or link to this item:
Title: Machine learning and simulation of GPR data
Authors: Tan, Jia Dian
Keywords: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
Issue Date: 2022
Publisher: Nanyang Technological University
Source: Tan, J. D. (2022). Machine learning and simulation of GPR data. Final Year Project (FYP), Nanyang Technological University, Singapore.
Project: B3110-211
Abstract: Ground penetrating radar (GPR) is a geophysical inspection method that makes use of electromagnetic waves to detect objects underneath a surface. In this project, a center frequency GPR as well as a dual polarizing GPR were used to carry out underground tree root detection by obtaining image scans, mainly in the form of b-scans. Currently, there are limitations to the number and nature of b-scans that can be obtained from experimentation. Not many ground truth b-scans of tree roots are readily available from the GPRs that the team is using. As such, a machine learning approach to image generation was explored to generate b-scans that can reproduce realistic and novel representations of these scans. In this paper, the deep learning domain of Generative Adversarial Network (GAN) will be explored and implemented to achieve the generation of the b-scans. Drawbacks and improvements to the model will also be explored, such as using other variants of GANs through Deep Convolutional GANs and Wasserstein GANs. Other uses of GANs will also be explored to complement the generation of realistic b-scans. One such usage will be the image translation capability of GANs to add realistic features to ground truth b-scans to obtain the desired b-scans for this project.
Schools: School of Electrical and Electronic Engineering 
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
2.84 MBAdobe PDFView/Open

Page view(s)

Updated on Jun 21, 2024

Download(s) 50

Updated on Jun 21, 2024

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.