A Deep Learning framework for Ground Penetrating Radar

Ourania Patsia, Antonios Giannopoulos, Iraklis Giannakis

Research output: Chapter in Book/Report/Conference proceedingPublished conference contribution

Abstract

Machine learning (ML) is becoming a more frequently used approach to deal with GPR and other electromagnetic problems, which due to the complexity of the data, require new more complex solutions. We have developed an ML framework to provide solutions to specific GPR applications and scenarios. The ML tools utilize neural networks (NNs) and a large training set originating from simulations that include a digital twin of a real GPR transducer. The applications investigated are background removal, automatic estimation of the background bulk permittivity in conjunction with a reverse time migration (RTM) scheme that utilizes the ML outputs and is applied to reinforced concrete slab scenarios. The schemes are validated using both synthetic and real data, showing a very good accuracy and demonstrating the success of the ML algorithms. Although, this ML framework is applicable to certain applications and scenarios, it can be easily extended to other classes of problems.

Original languageEnglish
Title of host publication2021 11th International Workshop on Advanced Ground Penetrating Radar, IWAGPR 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665422536
DOIs
Publication statusPublished - 2021
Event11th International Workshop on Advanced Ground Penetrating Radar, IWAGPR 2021 - Valletta, Malta
Duration: 1 Dec 20214 Dec 2021

Conference

Conference11th International Workshop on Advanced Ground Penetrating Radar, IWAGPR 2021
Country/TerritoryMalta
CityValletta
Period1/12/214/12/21

Keywords

  • Training
  • Deep learning
  • Maximum likelihood estimation
  • Ground penetrating radar
  • transducers
  • digital twins

Fingerprint

Dive into the research topics of 'A Deep Learning framework for Ground Penetrating Radar'. Together they form a unique fingerprint.

Cite this