Enhanced imaging of concealed defects behind concrete linings using Residual Channel attention network for rebar clutter suppression.

Ground Penetrating Radar (GPR) is a geophysical tool widely used for near-surface detection and imaging, but its effectiveness in identifying defects behind reinforced concrete linings is often reduced by electromagnetic shielding and rebar clutter. This paper presents a systematic workflow that enhances defect echo detection by suppressing rebar clutter. A synthetic dataset, specifically tailored to double-layer rebar meshes, is constructed to develop an optimized Residual Channel Attention Network (RCAN) for effective rebar clutter suppression. Data processed through this network undergoes imaging via a Reverse Time Migration (RTM) algorithm enhanced by Poynting vector decomposition. Analyses of synthetic data highlight the impact of rebar mesh on concealed defects and validate the approach presented. Field validations on full-scale tunnel linings confirm the algorithm's robustness and applicability, demonstrating its potential for accurately imaging concealed defects behind reinforced concrete linings. [Display omitted] • Systematic workflow for precise imaging of concealed defects behind reinforced concrete linings. • An RCAN is established for suppression of rebar clutter in GPR data. • A comprehensive dataset, comprising pairs of synthetic GPR data, is established for training the network. • Enhancement of the RTM algorithm for high-quality GPR Imaging using Poynting vector decomposition. • Field experiments validate the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]