2023 Volume 35 Issue 1 Pages 561-566
Each element on the periodic table has its own unique peak spectrum in LIBS analysis. Therefore, the chemical composition of a sample can be determined by identifying the wavelength of each peak from the analyzed sample. This has been a common sample identification method. However, due to various measurement conditions and environments, this peak search with database reference is a very tedious task. In this study, we propose a new method of analyzing LIBS spectra using only standard samples and no database reference or peak wavelength search, based on a machine learning method.