Volume quantification in interphase voxels of ore minerals using 3D imaging

This contribution presents and validates a new method to correct for the main limitations of volume quantification using X-ray computed tomography: limited spatial resolution and lack of mineralogical classification. The volume of a phase of interest (cassiterite, SnO2) is calculated using the intensity of voxels at interphases, which are typically the regions of main uncertainty in 3D imaging. Instead of traditional segmentation methods that define boundaries between phases, our method considers interphases as regions that can be several voxels across. The method is validated using a feedback loop between 2D scanning electron microscopy-based image analysis and bulk chemical analysis where the advantages of each technique are used to correct for the limitations of another. The percent of cassiterite derived from our method are within 10% deviation from those measured by scanning electron microscopy-based image analysis and bulk chemical analysis, when the P50 of the particle size distribution is at least 5 times the voxel size of the scan, which is a better agreement than results derived from other segmentation methods. Therefore, our method reduces the uncertainty of volume quantification and lowers the limit of grain sizes for which volumes can be reliably measured using computed tomography. The reduced uncertainty and bias can contribute to broadening the application of 3D imaging to mineral engineering as complementary to well established techniques.