Suitability of structure from motion for rock‐slope assessment

Abstract

Abstract This study examines three sites in Alaska with unstable rock slopes that were surveyed using both terrestrial laser scanning ( TLS ) and structure‐from‐motion (SfM) techniques. The datasets were acquired simultaneously and linked to a rigorous survey control network. An accuracy evaluation of the SfM‐derived surface models was performed using the TLS data and numerous reflectorless total station observations collected across the rock slopes. A quality evaluation was conducted to examine differences in point density, model completeness and distributions of morphological properties between the SfM and TLS datasets. The results indicate that SfM is a viable option for unstable rock‐slope assessment when a sufficient number of images with adequate overlap are acquired and the reconstruction is tied to a survey control network. The best results, in terms of accuracy and completeness, were achieved when combining both unmanned aerial vehicle ( UAV ) and terrestrial imagery for the SfM reconstruction. However, issues such as over‐smoothing and geometric inconsistencies bring into question the suitability of SfM for the detection of small changes over time.