Semi-Automated rock block volume extraction from high-resolution 3D point clouds for enhanced rockfall hazard analysis

Abstract

Rockfalls are critical landslide phenomena affecting human activities, with risk assessment based on hazard evaluation and potential impacts on exposed elements. Traditional methods for estimating unstable rock block volumes require direct measures often in hard-to-reach areas, hazardous, with time consuming approaches. This study introduces a semi-automatic method for estimating the most probable volume of the unstable blocks using open-source software (CloudCompare) to process 3D Point Cloud (PC) data obtained via Terrestrial Laser Scanning (TLS). The application area is a rock slope in a coastal sector of the northern Sicily (Italy) affected by frequent rockfalls phenomena. Both traditional field surveys and TLS were employed to characterize discontinuities and perform kinematic analyses. Volumes of already fallen blocks were directly measured, while unstable blocks were identified and volumetrically assessed using the PC-based procedure. Statistical analysis revealed that both created datasets conform to lognormal distributions; direct measurements show a better fit due to a larger sample size. Moreover, direct and indirect approaches were applied for recognition of main discontinuity sets influencing block detachment through planar sliding, toppling, and wedge failure. The proposed method offers a safer, more efficient alternative for rock mass characterization. Integration of traditional and remote sensing techniques facilitates accurate hazard evaluation, enhancing risk reduction strategies in vulnerable areas.