Regresar

Digitalisation of rock specimens and outcrops for training

Abstract:

Three types or classes of rocks are commonly used for rock classification according to its genesis: sedimentary, igneous and metamorphic. Their classification is based on some characteristics of rocks such as composition, texture, strength, etc. Experienced users can identify rocks observing certain features (colour, grain size, foliation, presence of fossils, etc.), but non-experts may require training to reach this competence. Several resources are available on the Internet and on the scientific and teaching literature presenting photos of rocky specimens. However, the experience of handling a rock on hands not only provides more information of the specimen but also makes the main characteristics of the rock easier to remember. The form of an outcrop depends, among other factors, on the lithology (i.e., presence of bedding planes, weathering, discontinuities, etc.). In both cases, digital 3D models of rock specimens and rocky slopes aid the users to train their abilities and recognition skills. We present a benchmark to aid the training process that classifies the rocks and outcrops using a textural classification. It allows inspecting the 3D textured models of rocks and slopes along with a behavioural classification. Two techniques were used to generate the models: 3D laser scanning (3DLS) and structure from motion (SfM). 3DLS-derived datasets are a 3D point cloud used to generate a 3D mesh over which the texture is superposed from the digital photos. To generate the SfM datasets we used digital photos, and therefore the process incorporates the texture. The benchmark is open access and collaborative in the URL: https://web.ua.es/digitalrocks, and it is being used by geology and engineering students for their training.