Answers are grounded in this paper and cite their sources.
Demonstrates improved cavity localization in urban environments through integrated methods.
Detecting strongly localized cavities in heterogeneous urban near-surface environments remains challenging for conventional surface-wave methods due to wavelength-averaged resolution limits. This study proposes an integrated workflow that combines multichannel analysis of surface waves (MASW), Rayleigh-wave back-scattering analysis, and Random Forest–based probability mapping to enhance cavity localization. The approach is evaluated at a controlled test site containing two engineered cavities with known geometry. Results show that while MASW reliably constrains burial depth, it produces laterally smoothed anomalies for meter-scale cavities. In contrast, frequency-wavenumber-based isolation of reverse-propagating surface-wave components provides sharper lateral constraints. The machine-learning model further integrates multichannel waveform and spatial-correlation features to generate continuous cavity-probability profiles, consistently identifying cavity centroids in field data and demonstrating stable classification performance in independent synthetic tests. These findings demonstrate that combining dispersion-based inversion, scattering-sensitive localization, and data-driven probability estimation improves detection reliability for small, shallow cavities in complex backfill conditions.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: