Rock fractures analysis using Structure from Motion technology: new insight from Digital Outcrop Models

Fractures are one of the most important features of the rocks of the upper crust since they strongly influence their physical and chemical behavior and reflect their tectonic history. For this reason, fracture study plays a key role in different branches of the geosciences. Notwithstanding, the quantification of the features and parameters describing fractures could be unsatisfactory using the standard field techniques because they are mainly based on direct-contact methodologies that are affected by errors, as orientation bias and trace censoring, and scarce representativeness, due to the limited possibility of acquiring information of outcrops partially or totally inaccessible. Recently new remote sensing technologies, such as Terrestrial Laser Scanner (TLS) and Digital Photogrammetry (DP), can help to overcome these limitations. Whereas TLS could be very expensive and difficult to use in geological study, DP permits to obtain similar results in an easier way due to cheaper and lighter equipment and more straightforward procedures. Moreover, DP becomes even more useful when combined with Unmanned Aerial Vehicle (UAV) because permits to acquire digital images from positions inaccessible to humans, allowing to analyze geological objects from points of view previously unimaginable. The images acquired from the ground and/or by the UAV can be then processed using different digital algorithms, such as Structure from Motion (SfM), that permit to create 3D model of the studied outcrop. In geosciences, the 3D model representing the surface of the outcrop is often called Digital Outcrop Model (DOM). Despite DOMs can be really useful in different branches of geosciences, their applications are quite well limited because the procedures of their development and sampling/analysis are scarcely analyzed in literature. It is important to highlight that whereas the UAV-based SfM approach is fairly discussed in literature for simple flat areas, is scarcely treated for application to near vertical and not-planar slopes. Moreover, the validity of some procedures of fracture sampling on 3D model, with special regards to the automatic ones, that have been recently presented in literature, is not well treated for real cases of study. The scarce knowledge about these approaches could cause different troubles to the scientific-users: from the application of avoidable time-consuming routine, to the acquisition and interpretation of erroneous data. This research aims to contribute to the scientific knowledge of the use of digital photogrammetry for fractured rock mass analysis, creating and defining new approaches and procedures for the development, analysis and application of DOMs. Here, a workflow for the fracture analysis of steep rocky outcrops and slopes using the 3D DOM is presented. In particular, the following steps are discussed: (i) image acquisition; (ii) development of 3D model; (iii) sampling of DOM; (iv) quantification and parametrization of the 3D measures; (v) application of the 3D quantitative data and parameters to different case of study. Four different cases of study were selected to validate the proposed method: the upper Staffora Valley and Ponte Organasco (Northern Apennines, Italy), Ormea (Ligurian Alps, Italy), and Gallivaggio (Western Alps, Italy) cases of study. However, this methodology could not completely replace the 'direct-contact' field activity, because some information as roughness, infilling and aperture of fractures cannot be measured satisfactory, and because, where possible, field control measures to validate the 3D data are necessary. However, this methodology could be considered as a new necessary procedure for rock-fracture studies because it allows to overcome the inevitable errors of the ground-based traditional methodology and because the DOMs are always available for the analysis, promoting data sharing and comparison, two fundamental principles on which science have and will have to be based