Integration of robotic total station and digital image correlation to assess the three-dimensional surface kinematics of a landslide

Abstract

In the field of landslide monitoring, the assessment of the spatially-distributed three-dimensional surface displacement is crucial to understand the underlying mechanisms. Nevertheless, available technologies and techniques that provide such a datum are few and often suffer spatio-temporal resolution, logistic and/or financial limitations. In this framework, we developed a methodology that merges the three-dimensional measurements at specific points, acquired by a robotic total station (RTS), and the spatially-distributed data obtained with digital image correlation (DIC) of time-lapse camera photographs, to achieve the spatially-distributed three-dimensional surface displacement. The integration method follows this procedure: i) the DIC results are orthorectified on an existing digital elevation model; ii) the RTS data are rototranslated into the camera coordinate system; iii) the ratio α between displacement vertical component and module measured by the RTS is calculated and interpolated across the region of interest; iv) the orthorectified DIC results are rescaled according to α, obtaining the three surface displacement components; v) the displacement vector is rototranslated into the geographical coordinate system. The sensitivity analysis respect to α revealed that the integration method can be successfully applied even with a limited number of RTS measurement points. The developed methodology has been applied to the Mont de La Saxe rockslide case study, during a phase of strong acceleration. In this period, the displacement magnitudes varied between 0.1 m and 10 m, thus providing a stress-test input for methodology development and validation. The results have been compared with independent ground-based interferometric radar measurements, obtaining 0.99 linear correlation coefficient and median absolute deviation of 0.086 m, which is comparable with the DIC measurement uncertainty. The proposed method is based on the use of low-cost portable and commonly used field equipment, thus it can be easily implemented in existing monitoring networks without additional financial costs.