EUVIMEDEUVIMED entdecken
Crossref

Improving ground penetrating radar image classification using deep learning with white noise data augmentation

Quý Hoàng Ngọc, Seonghun Kang, Sub Lee Jong, Hùng Trương Quang

Transport and Communications Science Journal · 12.12.2024 · 76 (6) · 916-927

DOI: 10.47869/tcsj.76.6.8 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Transport and Communications Science Journal, online erschienen am 12.12.2024. Verfasst von Quý Hoàng Ngọc und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Ground Penetrating Radar (GPR) technology is increasingly utilized in underground structure surveys in Vietnam to enhance construction safety, optimize costs, and minimize environmental impact, particularly in urban infrastructure, transportation, water resource management, and archaeological studies. However, to maximize its potential, advancements in data processing and analytical techniques are required to improve the accuracy of subsurface investigations and geological assessments. This study investigates the enhancement of GPR image classification performance through the integration of deep learning techniques (DNN) and data augmentation utilizing white noise. The dataset used in the research was collected from construction sites in South Korea and comprises three primary categories: background noise, hyperbolic reflections from buried objects, and reflections from manholes. To increase the diversity of the dataset and mitigate the effects of environmental noise, a data augmentation method incorporating white noise with a coefficient of variation (cov) is employed. A DNN model based on the VGG19 architecture is implemented for feature extraction and classification of GPR images. Experimental results demonstrate that the augmented data significantly enhance classification performance, reduce overfitting, and improve the generalization capability of the model. This research highlights the potential of using white noise-based data augmentation to enhance GPR image classification performance

Zitieren

APA

Ngọc, Q. H., Kang, S., Jong, S. L., & Quang, H. T. (2024). Improving ground penetrating radar image classification using deep learning with white noise data augmentation. Transport and Communications Science Journal, 76(6), 916-927. https://doi.org/10.47869/tcsj.76.6.8

Vancouver

Ngọc QH, Kang S, Jong SL, Quang HT. Improving ground penetrating radar image classification using deep learning with white noise data augmentation. Transport and Communications Science Journal. 2024;76(6):916-927. doi:10.47869/tcsj.76.6.8

BibTeX
@article{Ngc2024,
  author = {Ngọc, Quý Hoàng and Kang, Seonghun and Jong, Sub Lee and Quang, Hùng Trương},
  title = {Improving ground penetrating radar image classification using deep learning with white noise data augmentation},
  journal = {Transport and Communications Science Journal},
  year = {2024},
  volume = {76},
  number = {6},
  pages = {916--927},
  doi = {10.47869/tcsj.76.6.8}
}
Kostenlos starten →