Digital 3D Rocks: A Collaborative Benchmark for Learning Rocks Recognition

Abstract

Naked eye rock recognition is an essential activity for professionals and students of geosciences, architecture and engineering. Through a hand holding rock specimen, it is usually required not only to identify the type of rock but recognize their texture and understand its expected properties mechanical and petrophysical properties. Although a wide choice of books, websites and apps are available in the literature and on the Internet, their contents are two-dimensional (2D) and static. Nowadays, the application of remote sensing techniques such as Light Detection and Ranging (LiDAR) or Structure from Motion (SfM) enable the generation of three-dimensional (3D) interactive models, which are here presented as a novel perspective of learning and practising rocks recognition. Despite limitations of the technique, 3D digital models of rocks permit their virtual visualization and manipulation to reveal parts of the specimens that are hidden in the 2D photograph, as well as details of the rock specimen’s texture such as grain and minerals size, distribution and organization along with the possibility of identifying petrological features, foliation, mineral orientations and others. This provides a novel perspective of learning and practising rocks identification. Herein, a benchmark of digital rocks collected all around the world and generated using SfM technique is presented. The rocks are organised using a straightforward classification system based on the texture jointly with a detailed description to aid the specimen recognition. A behavioural geomechanical classification is then applied. Moreover, a linked datasheet shows the engineering classification, the weathering degree, the guide physical and mechanical properties (general, and specific when available), the engineering uses and others. The information is organised on an open-access website hosted by the University of Alicante (https://web.ua.es/digitalrocks). This initiative also aims to encourage students and professionals to generate their own models and to provide the description to enlarge the repository.