Drone-derived SfM Photogrammetry and Digital Rock Mass Mapping of Rock Slopes

Abstract

Rockfalls occur frequently in Norway and pose a risk to people and infrastructure. To determine suitable mitigation measures that could prevent fatalities and damages, rockfall hazard needs to be assessed based on representative and reliable data. The aim of this thesis was to test if drone-derived SfM photogrammetry and digital mapping are methods that can facilitate, streamline, and improve the determination of relevant parameters for rockfall hazard assessment. The method was applied to five study areas located in different parts of Norway, each containing a natural rock slope susceptible to rockfalls. The objectives of this thesis were to establish a functional survey design for collecting data of sufficient quality, establish an efficient workflow for generating and validating georeferenced 3D models, and evaluate different tools for digital mapping of rock masses. During fieldwork, datasets of overlapping images were collected with a drone-based consumer-grade camera and ground control point positions were recorded with GNSS receivers. Additional GCPs were extracted from orthophotos and digital elevation models. Photogrammetric processing of the collected datasets and generation of 3D models were performed using Agisoft Metashape. Cloud-to-cloud comparison was used to validate the model quality. The model quality was validated with cloud-to-cloud comparison, using high-resolution LiDAR models as reference. Digital mapping of discontinuities was performed on the validated 3D models, using Maptek PointStudio. Discontinuities were characterized by orientation, spacing, persistence, and roughness. Kinematic analyses were performed on the digitally mapped discontinuities and likely failure modes were determined. The established survey design and workflows were found to be functional, efficient, and capable of generating reliable data that accurately represents the study areas.