Automated digital twin reconstruction for tunnel inspection and maintenance

Routine inspections of underground structures are essential for safety. However the inspection process is labour-intensive and evaluations can be subjective, highlighting need for automation. While BIM has improved digital management, its current adoptions has not fully exploited the digital revolution. Recent large models like Segment Anything Model (SAM) shows great potential in automation, enabling object recognition without training. This paper proposes a framework for reconstruction of Digital Twins (DTs) using BIM technology, focusing on automation in component detection, digital model reconstruction, condition mapping and quantification. Tunnel point clouds are converted into images and processed by SAM for effective zero-shot segmentation, outperforming supervised learning. Then, geometric parameters were extracted from previous results for reconstructing DTs, deformation and displacement monitoring. The fine-tuned SAM also used for defect detection and recorded in digital model, providing automated assessments of defect severity. This framework enhances tunnel maintenance through remote inspection, data integration, and automated assessments.