Ground penetrating radar (GPR) is one of the most recommended tools for routine inspection of tunnel linings. However, the rebars in the reinforced concrete produce a strong shielding effect on the electromagnetic waves, which may hinder the interpretation of GPR data. In this work, we proposed a method to improve the identification of tunnel lining voids by designing a generative adversarial network-based rebar clutter elimination network (RCE-GAN). The designed network has two sets of generators and discriminators, and by introducing the cycle-consistency loss, the network is capable of learning high-level features between unpaired GPR images. In addition, an attention module and a dilation center part were designed in the network to improve the network performance. Validation of the proposed method was conducted on both synthetic and real-world GPR images, collected from the implementation of finite-difference time-domain (FDTD) simulations and a controlled physical model experiment, respectively. The results demonstrate that the proposed method is promising for its lower demand on the training dataset and the improvement in the identification of tunnel lining voids.

Ground penetrating radar (GPR) is considered an effective tool to detect tunnel lining voids. In this paper, an experimental study was carried out using a physical tunnel lining model to evaluate the performances of different antenna frequencies. We built a 4.2 m–long, 4.2 m–wide, and 2.0 m–high experimental model to simulate the secondary lining, initial lining, and surrounding rock of a tunnel structure. In the model, we created four categories of voids, which are voids in secondary and initial linings, a delamination between the secondary and initial linings, a delamination between the initial lining and sand, and a void buried in the sand, to simulate real cases in tunnel engineering. The GPR wave velocities in the sand and concrete of the model were measured using the reflection method for the calibration of void depth. We employed a commercial GPR system equipped with antennae of different centre frequencies to detect the voids. GPR data were processed using a conventional data processing flow, and the performances of different frequencies were examined. The results show that the 1000 MHz centre frequency GPR is capable of characterizing shallow buried voids in the secondary lining but is not able to penetrate into the initial lining. The 250 MHz centre frequency GPR system is not advised to detect voids in or behind tunnel linings due to its low resolving power for voids of centimetre sizes. The 500 MHz centre frequency GPR system is optimal for void detection because it demonstrated a balanced performance of resolving ability and investigation depth. The findings of this work could be useful references for antenna selection and data processing in real GPR applications.