A ROS 2-enabled VIS reflectance and UV-induced fluorescence mini-spectrometer for lithology recognition by mining robot prototypes.

A ROS 2-enabled VIS reflectance and UV-induced fluorescence mini-spectrometer for lithology recognition by mining robot prototypes This archive contains the data, analysis code, firmware and hardware design files supporting the paper "A ROS 2-enabled VIS reflectance and UV-induced fluorescence mini-spectrometer for lithology recognition by mining robot prototypes" (submitted to Computers & Geosciences). The module is a low-cost (< EUR 500) VIS reflectance and UV-induced fluorescence spectrometer built around a Hamamatsu C12880MA sensor (288 channels, ~315–880 nm), a high-CRI white LED (Nichia NF2W757GT-F1 Optisolis, CRI 98) and a 365 nm UV LED. It publishes spectra directly on a ROS 2 bus and was evaluated on drill cuttings from seven boreholes in the Brussels-Capital Region (Belgium). About 27,000 spectra were acquired across 760 depth intervals. On cuttings from a borehole held out of training, a random forest classifier correctly assigns one of eight stratigraphic classes 66% of the time; adding a stratigraphic-order constraint raises this to 76%. Contents data/ — labeled spectra from the seven boreholes (dry and wet), interval logs, class mapping, and raw per-interval measurementscode/ — Python pipeline used to generate the paper's programmatically produced figuresfigures/ — output directory populated by the analysis pipelineresults/ — result tables and per-interval prédictionshardware/ — Eagle schematic and board files, 3D housing (STEP/STL/F3D), principle schematic, LED emission spectrafirmware/ — standalone Arduino firmware, RM3 micro-ROS firmware, and RM1 ROS 2 driversoftware/ — PyQt5 acquisition GUI used during the borehole campaigns Data summary Seven boreholes from two anonymized drilling sites: site 1 (deep geothermal boreholes F3, F4) and site 2 (geotechnical/hydrogeological boreholes P1, P2, P3, P4, PZS), roughly 4 km apart. Each depth interval was measured ten times in reflectance and ten times in fluorescence, dry and, where available, wet. Intervals are labeled with the contractor log's lithology, harmonized into eight classes. Reproducing the analysis bash python -m venv .venv && . .venv/bin/activate pip install -r requirements.txt cd code python run_all.py This regenerates the paper's programmatically produced figures and result tables. Individual pipeline steps can also be run on their own using the shipped prediction files. License Code, firmware and hardware files are released under the MIT License. Data and figures are released under CC BY 4.0. Funding This work was funded by the European Union's Horizon 2020 research and innovation programme under grant agreement no. 820971 (ROBOMINERS, https://robominers.eu). The Brussels borehole dataset was made available through the BruGeo project (Brussels Environment / ERDF). The geological background map is served by Brussels Environment (Brugeo, BruStrati3D).