Slope-Scale Rockfall Susceptibility Modeling as a 3D Computer Vision Problem

Rockfall constitutes a major threat to the safety and sustainability of transport corridors bordered by rocky cliffs. This research introduces new approach rockfall susceptibility modeling for identification potential source zones. is achieved developing data-driven model assess local slope morphological attributes with respect rock evolution processes. The ability address “where” more likely occur via analysis historical event inventories terrain define probability given area producing critical advance toward effective corridor management. availability high-quality digital volumetric change detection products permits developments in assessment prediction. We explore simulating conceptualization slope-scale using computer power artificial intelligence (AI). employ advanced 3D vision algorithms analyzing point clouds interpret high-resolution observations capturing long-term, LiDAR-based differencing. has been developed tested on data from three slopes: two Canada one UK. results indicate clear AI advances develop indicators geometry learning recent activity. resultant models produce slope-wide maps high resolution, up 75% agreement validated occurrences.