Standoff LIBS data package from the ROBOMINERS project prototype

Standoff LIBS hyperspectral imaging of the Mežica Pb–Zn deposit (Slovenia): data, instrument and classification code This package contains the data, acquisition/processing code and instrument documentation underpinning our study on standoff Laser-Induced Breakdown Spectroscopy (LIBS) hyperspectral imaging for robotic mining perception, developed within the EU H2020 ROBOMINERS project (Grant No. 820971). LIBS hypercubes were acquired on Mississippi-Valley-type (MVT) Pb–Zn mineralisation at the Mežica tourist mine (Slovenia) and classified pixel-by-pixel into mineral/ground-truth classes (sphalerite, galena, dolomite, calcite, iron oxide/oxyhydroxide). Classifications were independently validated by micro-Raman spectroscopy and SEM-EDS on hand samples. Contents LIBS hypercubes/ — Standoff LIBS spectral image cubes in NetCDF (.nc, dimensions bands × y × x, dual-spectrometer, ~4096 channels) for Mežica tourist-mine Sites 1–3, plus focus-calibration test cubes (flat and angled targets). The Site 3 folder also includes the labelled train/test sets (.json), classification reports (REPORT *.txt), the training/testing set composition (.docx), and co-registered white-light and UV reference images (.tif). Hypercube processor (classification) script/ — hypercube_processor.py (with help.html): Python tool to load the .nc cubes, preprocess the spectra and run the pixel classification. StandoFf LIBS instrument/ — Instrument documentation and control code: README.md, mechanical CAD of the standoff head (hardware/*.step, .png), and acquisition scripts (mapping_script_standoff.py — gated, hardware-triggered point-by-point mapping; lens_calculator_app.py; laser_heating_routine.py) plus a LIBS emission-line reference list (rbins_libs_bands.csv). Supplementary_S2_raw_data/ — Validation analyses (Supplementary Data S2): Raman/ — raw Bruker OPUS files, baseline-input sample spectra (.csv) and matched RRUFF reference spectra (.csv) for seven analysed spots. SEM-EDS/ — point-analysis summary (EDS_point_analyses.csv) and raw AZtec exports (.xlsx) for sphalerite, galena, dolomite and iron-oxyhydroxide. Formats and usage NetCDF cubes can be opened with xarray/netCDF4 (Python) or Panoply; spectra are plain CSV; spectrometer raw data are in Bruker OPUS format; instrument hardware is provided as STEP CAD. See the instrument README.md and the classifier help.html for details.