Machine learning for SMART mineral mapping using coupled XRF-XRD скачать в хорошем качестве

Machine learning for SMART mineral mapping using coupled XRF-XRD

A Synchrotron-based Machine learning Approach for RasTer (SMART) mineral mapping was developed to train a mineral classifier that can interpret raster-scanned millimeter-sized areas of rock thin sections with micron-sized resolution. Training is done using Artificial Neural Networks (ANN) with coupled micro X-ray fluorescence (µXRF) intensities, which provide information about element abundances, and micro X-ray diffraction (µXRD) patterns, which provide information about mineral identity. The resulting SMART mineral mapper can identify minerals using only micro X-ray Fluorescence (µXRF) data. The value of this approach comes from the fact that µXRF data are relatively fast to collect and interpret whereas the µXRD data take longer to collect and much longer to interpret. "SMART mineral mapping: Synchrotron-based machine learning approach for 2D characterization with coupled micro XRF-XRD" Julie J. Kim, Florence T. Ling, Dan A. Plattenberger, Andres F. Clarens, Antonio Lanzirotti, Matthew Newville, Catherine A. Peters Computers & Geosciences Volume 156, 2021, 104898, ISSN 0098-3004, https://doi.org/10.1016/j.cageo.2021..... *Corresponding author. Dr. Catherine A. Peters ([email protected]), Department of Civil & Environmental Engineering, Princeton University, Princeton, New Jersey, 08544 Presented at Goldschmidt Virtual 2020 Data and computer code are available and accessible on Digital Rocks Portal: Peters, C.A., & Kim, J.J. (2020). “Eagle Ford Shale: Synchrotron-Based Element and Mineral Maps”. https://www.digitalrocksportal.org/pr... DOI: 10.17612/T3A6-6356 https://doi.org/10.17612/T3A6-6356