A Sliding Time‐Window Energy Ratio Method for Weak Signal Enhancement of Deep Subgrade Voids in GPR Data

ABSTRACT

Deep subgrade voids and interfacial debonding in pavement often produce weak ground‐penetrating radar (GPR) responses due to propagation attenuation and interference from shallow layers. Conventional gain compensation and whole‐trace energy methods struggle to isolate these concealed targets, as the weak deep signals are easily masked by background clutter. To address this, a sliding time‐window energy ratio (STWER) method is proposed to enhance deep weak reflections. This approach effectively transforms weak reflections into localized energy‐deviation features. The method was validated using gprMax numerical models under varying burial depths, sizes, and dielectric properties, alongside field GPR data (900 MHz and 270 MHz) verified by excavation and core drilling. Numerical results demonstrate that STWER consistently highlights weak void responses despite depth‐induced attenuation, generating stronger energy‐ratio peaks for air‐filled voids compared to non‐void dielectric anomalies. Field case studies further confirm that STWER provides more targeted anomaly localization than automatic gain control and global energy ratios by minimizing non‐target temporal interference. Consequently, STWER serves as an efficient auxiliary technique for the rapid identification of deep pavement defects.