Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces the "Minerals in the Wild" dataset, which includes 1,132 rock specimens with hyperspectral imaging (HSI) data and elemental characterization from X-ray fluorescence (XRF) sensors. The dataset addresses the lack of publicly available, reliable ground-truth labels for mineral identification, which is crucial for advancing hyperspectral imaging techniques in mineral exploration and processing. The authors also propose a pruning mechanism to enhance the matching of HSI pixels to USGS spectral signatures, demonstrating improved performance over simpler baseline methods.
A new dataset with over 1,100 rock specimens reveals that targeted pruning of spectral signatures can significantly boost the accuracy of mineral identification using hyperspectral imaging.
Rapid mineral characterization is essential for applications ranging from mineral exploration to industrial ore processing. To this end, Hyperspectral Imaging (HSI) has emerged as a promising sensing modality thanks to its fine spectral resolution, enabling mineral discrimination in both close-range and remote sensing settings. However, the scarcity of publicly available datasets with reliable ground-truth labels hinders the development and evaluation of HSI-based mineral identification methods. We release Minerals in the Wild, a multi-purpose dataset comprising 1,132 rock specimens collected across Europe. For each specimen, we provide an HSI acquisition together with an elemental characterization obtained via an XRF sensor. We define the task of elemental characterization on our dataset and propose a pruning mechanism that removes distant signatures from the USGS dictionary prior to a convex optimization approach for matching HSI pixels with USGS spectral signatures. Finally, we empirically show that our approach outperforms simpler baselines.