Training-Free Tunnel Defect Inspection and Engineering Interpretation via Visual Recalibration and Entity Reconstruction

TL;DR

TunnelMIND is a training-free framework that enhances tunnel defect inspection by recalibrating defect proposals and reconstructing structured defect entities for better engineering assessment.

Contribution

It introduces a novel training-free approach that transforms coarse defect proposals into detailed, structured defect evidence suitable for engineering documentation.

Findings

Achieves F1 scores of 0.68, 0.78, and 0.72 on visible, GPR, and road defect tasks.

Recalibrates spatial support of defect proposals using dense visual consistency.

Reconstructs defect entities with detailed attributes for engineering report generation.

Abstract

Tunnel inspection requires outputs that can support defect localization, measurement, severity grading, and engineering documentation. Existing training-free foundation-model pipelines usually stop at coarse open-vocabulary proposals, which are difficult to use directly in interference-heavy tunnel scenes. We propose a training-free framework TunnelMIND. Specifically, language-guided defect proposals are not treated as final outputs; instead, their spatial support is recalibrated at inference time through dense visual consistency, so that coarse semantic anchors can be transformed into more reliable prompts under tunnel-specific hard negatives. The resulting masks are further reconstructed into structured defect entities with category, location, geometry, severity, and context attributes, which are then mapped to retrieval-grounded explanation and engineering-readable report generation under expert knowledge constraints. On visible, GPR, and road defect tasks, TunnelMIND achieves F1 scores of 0.68, 0.78, and 0.72, respectively. Overall, TunnelMIND shows that training-free tunnel inspection can move beyond coarse localization toward structured defect evidence for engineering assessment.