The classification and discrimination of minerals are important in geological research. The distribution of rocks and minerals may be inferred based on their identification, which is helpful for the investigation of s...The classification and discrimination of minerals are important in geological research. The distribution of rocks and minerals may be inferred based on their identification, which is helpful for the investigation of some geological and environmental evolution problems, either on the earth or on other planets. LIBS and Raman spectra techniques have shown great advantages in simultaneous multi-component, in-situ, and non-destructive detection and they play an important role in rock composition analysis. In this presentation, six kinds of minerals （Gypsum, Spodumene, Barite, Haematite, Moonstone, and Labradorite） were detected by first using LIBS and Raman systems, and then several other methods （PCA, PLS-DA, ANN, and SVM） were used to evaluate the LIBS, Raman, and the fused LIBS/Raman data, respectively. The results indicate the superiority of the fused LIBS/Raman data in mineral classification, which stems from their complementary analysis abilities when studying element composition and structural features.展开更多
基金supported by Shandong University of China(No.2014ZQXM004)National Natural Science Foundation of China(No.41503063)
文摘The classification and discrimination of minerals are important in geological research. The distribution of rocks and minerals may be inferred based on their identification, which is helpful for the investigation of some geological and environmental evolution problems, either on the earth or on other planets. LIBS and Raman spectra techniques have shown great advantages in simultaneous multi-component, in-situ, and non-destructive detection and they play an important role in rock composition analysis. In this presentation, six kinds of minerals （Gypsum, Spodumene, Barite, Haematite, Moonstone, and Labradorite） were detected by first using LIBS and Raman systems, and then several other methods （PCA, PLS-DA, ANN, and SVM） were used to evaluate the LIBS, Raman, and the fused LIBS/Raman data, respectively. The results indicate the superiority of the fused LIBS/Raman data in mineral classification, which stems from their complementary analysis abilities when studying element composition and structural features.
