期刊
OPTIK
卷 158, 期 -, 页码 1058-1062出版社
ELSEVIER GMBH, URBAN & FISCHER VERLAG
DOI: 10.1016/j.ijleo.2017.12.167
关键词
Mineral recognition; LIES; The modified correlation analysis; Intensity ratios
类别
资金
- Natural Science Foundation of China [41503063]
- National Key Research and Development Program of China [2016YFC0302101]
The recognition of natural geological samples is of great significance for mineral exploration and geological research. Because of the powerful element analysis ability, the laser-induced breakdown spectroscopy data was widely used in the classification and recognition of sam- pies in recent two decades. Correlation coefficient, as a classical statistic parameter, reflects the degree of a close correlation between variables, and can be used to compare the LIBS data of test samples with the existing data. But it is difficult to get high accuracy in dealing with the identification problems of natural geological samples, which are usually composed of similar elements and mixed with minor impurity constituents. So. some typical spectral data segments and quantitative features, such as the intensity ratios of elements of LIBS data, are extracted and combined with the correlation coefficient by a specially designed analytical model to identify the geological samples. The modified correlation coefficient method shows a positive improvement in the recognition accuracy, while less complexity than the widely used pattern recognition methods. (C) 2018 Elsevier GmbH. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据