Reopening abandoned underground mines poses two significant issues: first, underground opening spatial information may not be current; secondly, sites are often flooded. A complex, often an uncharted network of flooded tunnels is a highly hazardous environment for traditional exploration methods, such as human diving. However, to make the right decision about reopening the mine, acquiring the geoscientific, topological, and structural data is necessary.  UNEXMIN GeoRobotics (UGR) was formed to answer these challenges, and for that purpose, it has built UX-series robots. The newest one, the UX-1Neo, is a semi-autonomous spherically shaped submersible with a 0.7 metre diameter and a 500 metre (50 bar) depth rating. Various sensors are installed on the UX-1Neo for mapping underground openings. Therefore, it can generate a detailed 3D map with a resolution within a few centimetres. Sensors installed include light projectors and multibeam and scanning sonars. Instrumentation installed for collecting geoscientific data includes a multispectral camera, a sub-bottom profiler, a water chemistry unit, and a water sampler. Furthermore, it has six cameras that provide a full 360 degrees high-quality perception of underground openings.  The UX-series robots have demonstrated its capabilities at seven different sites: Kaatiala pegmatite mine in Finland, Idrija mercury mine in Slovenia, Urgeiriça uranium mine in Portugal, Ecton Cu-Zn-Pb mine in the UK, Molnár János cave system in Hungary, Csór water well in Hungary, and Solotvyno salt mine in Ukraine. These field trials showed that the UX-series robots are capable of performing different tasks, such as surveying underwater structures, mineral exploration, and geological studies. Without human risks or environmental impacts, these robots can obtain high-quality mineralogical, geological, and topological data from inaccessible sites. The UX-2 is the next robot generation that is currently being developed. This robot has similar dimensions, but an increased depth rating of 1500 metres and a rock sampler unit; and will be used to address even more challenging scenarios.  Acknowledgments UGR is supported by the UNEXUP project, which is funded by EIT RawMaterials, project number 19160.