Modelagem geológica 3D utilizando aerofotogrametria com vant em mina de mármore no município de Caçapava do Sul – RS

Abstract

The increase in the demand for mineral goods has been demanding from mining companies the development of new methods for the exploration of these resources. The stripping ratio, or waste/ore ratio, determines the economic viability of a mining enterprise. The increase in stripping ratio implies the reduction of profits and/or the life of the mine. Geological modeling is an indispensable tool for mine planning. However, the construction of geological models requires time, planning, and expenditures. In a mine, often the rocks are located in inaccessible or hazardous areas. The use of photorealistic tridimensional models allows the remote classification of rocks occurring in a mine site, reducing the inherent risks and improving model quality. This work sought to apply an innovative methodology for the generation of a 3D geological model in a marble quarry, using a lithological classification based on remote sensing data. Unmanned aerial vehicles were employed for acquisition of aerial photographs, which were integrated by means of digital photogrammetry for the construction of 3D surface models. These models, denominated as virtual outcrop models, are composed of (i) millions of georeferenced points containing color and normal vector information (point clouds), and (ii) triangulated surface wireframes. The main lithologies occurring in the quarry were classified in the point cloud and the main exposed structures were interpreted virtually in the wireframe mesh. The data derived from the virtual outcrop models were processed with an open source machine learning algorithm. The result consists in a 3D lithological block model. From this block model it was possible infer the geological assembly up to 100 meters horizontally and 30 meters vertically from the terrain surface, which is enough to guide mine planning for at least two years.