Suitability of terrestrial archival imagery for SfM-MVS based surface reconstruction of steep rock walls for the detection of rockfalls

Abstract

Rockfalls are a common geomorphological process on steep rock slopes and therefore play an important role in the geomorphological dynamics of high mountain regions. Using multi-temporal high-resolution terrain models (DTMs) generated from remote sensing data from Light Detection and Ranging (LiDAR) and Structure-from-Motion (SfM) photogrammetry, surface changes can be calculated using a DTM of Difference (DoD). Whereas surface changes can usually only be covered over short investigation periods with terrestrial laser scanning (TLS) and digital photogrammetry, historical data sets enable surface reconstruction far back into the past. Therefore, our work focuses on the combined use of TLS data, terrestrial digital and terrestrial historical images to quantify rockfalls on the steep rock faces of the Tre Cime di Lavaredo in the Dolomites, Italy. The reconstruction of the topography of the rock faces from past times was preceded by an intense and time-consuming search for historical terrestrial photographs. After a complex data preparation and the development of a processing workflow adapted to the data, we were able to quantify the surface changes caused by rockfalls for the observation period 1970–2018. For two areas, we calculated surface changes of -101.28 m³ and -1,872.87 m³. Our study demonstrates the potential and limitations of historical terrestrial images for a long-term geomorphic change detection analysis of rockfalls. This specific workflow, depending on the quality of the images, has to be adapted for each dataset while constant controls of the intermediate results are indispensable.