A review of key technologies for ai-driven blasting excavation in drill-and-blast tunnels

Blasting excavation in drill-and-blast tunnels faces significant challenges, including delayed geological perception, experience-dependent drilling-and-blasting design, insufficient post-blasting quality feedback, and coupled safety risks. To promote the transition of blasting excavation from experience-driven to data-driven decision-making, this paper systematically reviews the application status of artificial intelligence technologies in blasting excavation of drill-and-blast tunnels based on recent domestic and international studies. First, the applications of machine vision, three-dimensional point clouds, and measurement-while-drilling technologies in intelligent geological perception are analyzed, and multi-source information processing and geological condition assessment are discussed. Second, research progress on machine learning in blasting vibration prediction, overbreak/underbreak identification, muck pile fragmentation evaluation, and multi-objective optimization of drilling-and-blasting parameters is summarized. Finally, the integrated application of digital twin technology in blasting excavation cycle modeling, monitoring feedback, and safety management is discussed. The results indicate that artificial intelligence has preliminarily supported the closed loop of “perception–prediction–optimization–feedback” in blasting excavation. However, challenges remain, including insufficient data standardization, limited cross-scenario generalization, and difficulties in multi-source heterogeneous data fusion. Future research should focus on edge-cloud collaborative real-time perception, post-blasting feedback-based parameter optimization, and reduced-manpower blasting excavation control under human-machine collaboration.