BRICK CLASSIFICATION BY COMBINATION OF PRINCIPAL COMPONENT ANALYSIS AND LINEAR DISCRIMINANT ANALYSIS USING LASER-INDUCED BREAKDOWN SPECTROSCOPY
| Authors | |
|---|---|
| Year of publication | 2014 |
| Type | Conference abstract |
| MU Faculty or unit | |
| Citation | |
| Description | Focusing on historical aspect, during archaeological excavation or restoration works of buildings or different structures built from bricks it is important to determine, preferably in-situ and in real-time, the locality of bricks origin. Combination of principal component analysis (PCA) and linear discriminant analysis (LDA) was applied on LIBS spectra, to classify altogether of 29 brick samples from 7 different localities. Comparison of two potentially field-deployable LIBS setups is introduced. The suggested methodology was tested on LIBS data obtained by stand-off and table-top setup. The table-top arrangement used for experiment was chosen with respect to possible construction of portable device. Lower pulse energy was used (as an alternative for diode-pumped laser) and CCD detector. Both setups provided ability of classification of bricks samples, but it was found out that database for training multivariate supervised method (LDA) has to be creatred by spectra measured on the same kind of device and using the same conditions as spectra to classify. The most notable novelty was the discover of bricks firing temperature effect on LIBS spectra revealed using PCA. This effect was closely investigated with the prospect of estimation of firing temperature of unknown samples. The dependence of firing temperature has very complex effect on spectra. Closer look showed that character of this dependence varies between classes of particular localities (where the reason can be in the mineralogical composition of clays, consequently in a degree of sintering or hardness of material) and also between two setups used (where the reasons can be in different plasma conditions or measuring conditions). |
| Related projects: |