Evaluation of Road Base Grouting Quality Using GPR and Image Recognition–Based Method

This study presents a case analysis of a road base grouting project utilizing three-dimensional (3D) ground-penetrating radar (GPR) and image recognition technology to evaluate grouting quality. Plane amplitude gray maps at 10 depths across nine cement concrete plates are analyzed before and after grouting, yielding 90 data sets. Key findings demonstrate (1) postgrouting grayscale maps exhibited significantly enhanced uniformity, with 94.4% of analyzed areas showing measurable improvement; this confirms grouting effectively homogenizes the road base structure; (2) gray value distributions transitioned from pregrouting bimodality to postgrouting unimodality, featuring reduced peak magnitudes and smoother profiles; this transition indicates substantially reduced dielectric contrast within the medium; and (3) gray value standard deviation exhibited strong correlation with grouting quality, establishing its utility as a reliable quantitative indicator. Conversely, mean values demonstrated limited diagnostic value for quality assessment. Validation via falling-weight deflectometer (FWD) retests and excavation confirmed the reliability of the proposed standard deviation-based index (δ). This methodology advances nondestructive evaluation in road maintenance, offering a practical framework for grouting quality assessment.