A deep learning network to improve tunnel lining defect identification using ground penetrating radar. Issue 4 (October 2021)

• Record Type: Journal Article
• Authors: Wang, Yuanzheng; Qin, Hui; Tang, Yu; Zhang, Donghao; Wang, Zhengzheng; Pan, Shengshan
• Is Part Of: IOP conference series. Volume 861:Issue 4(2021)
• Publication Date: 2021-10
• Subjects: Earth sciences -- Periodicals; Environmental sciences -- Congresses; Environmental sciences -- Periodicals; 550.5
• DOI: 10.1088/1755-1315/861/4/042057
• Languages: English
• ISSNs: 1755-1307

Abstract

Abstract: Ground penetrating radar (GPR) is one of the most recommended tools for routine inspection of tunnel linings. However, the rebar in reinforced concrete has strong shielding effect on electromagnetic waves, which makes the defects beneath the rebar hard to be distinguished in GPR images. To supress the reflection waves of the rebar network and reconstruct the defect echoes, we proposed a deep learning network based on generative adversarial networks (GAN). Taking the GPR images with rebar reflection waves as input, the network generated GPR images without rebar reflection waves. The GPR images processed by our networks are easier for manual interpretation and the existing object detection networks performs better on the images processed by proposed networks.