Currently, a geologist should manually fill an electronic table while he describes rocks or scratch a description on a blanket. This takes a lot of time and effort. Lots of researchers have developed their approach to autonomous rock classification. Unfortunately, all of them are based on a core column investigation. Most used a traditional approach to image classification based on a statistical analysis of images and machine learning techniques. Modern deep learning techniques were applied to classify and segment core images in this work. An important note is that previous works cover only core column examination, while this paper presents an approach to extract core columns from a core box. Thus, a geologist should not spend time on a manual description of the core or extraction of separate images out of a core box. This can be done by applying the developed algorithms described in this work. The algorithms were also included in a system which became worldwide available software.

This volume has its genesis in the “Backbone of the Americas: Patagonia to Alaska” conference that took place in the city of Mendoza, Argentina, from 2 to 7 April 2006. The conference was convened by the Asociación Geológica Argentina and the Geological Society of America in collaboration with the Sociedad Geológica de Chile and was the fi rst meeting on this scale to be jointly convened by North and South American geologic societies. Some 400 participants from 20 countries along the western margins of the Americas and elsewhere participated. The purpose of the meeting was to explore common scientifi c themes affecting the evolution of the Cordilleras of the Americas from a multidisciplinary international perspective. Major themes included collision of active and aseismic oceanic ridges, shallowing and steepening subduction zones, plateau and orogenic uplift, and terrane collision. This volume presents a selection of papers by the invited plenary speakers and other represenative presentations of the symposium <...>