Study Of Laser Induced Breakdown Spectroscopy Database And Classification For Geomateirai

Laser-induced breakdown spectroscopy (LIBS) is an emerging atomic emissionspectroscopy technique. There are many attributes especially the rapid, real-time andin-situ analysis with no sample preparation required that make this technique as anincreasingly attractive tool in geology field. Currently, LIBS applied for geologicalresearch is an emerging technology.Although there are many researches aboutgeomaterial analysis based on LIBS, a LIBS database for geomaterial with samplesdiscrimination capabilities has not been reported yet. In addtion, LIBS used forgeomaterial classification becomes a research focus in recent years, but a developedLIBS classification software has not been reported yet. LIBS database for geomaterialwith sample discrimination capabilities can provide a standard reference andmethodological support for geomaterial analysis based on this technique, which is anindispensable part of LIBS analysis for geomaterial and has an importantsignificance.The main research contentsare as follows:(1) Development of LIBS database for geomaterialLIBS database for geomaterial provides standard data and modelingbasis forgeomaterial element identification and sample classification using LIBS technique,respectively.LIBS database has been developed based on.NET framework using C#combined with MySQL server, which contains elements feature spectral database anda dedicated geological database. Elements feature spectral database includes2058lines for99elements (Z=1-99), and dedicated geological databasecontains1300LIBSemission lines for6rock reference species and7soil reference species.Databaseallows analysts edition of the database such as addition, deletion, correction andsearch of data. The developed database meets the design requirements, which can beused for LIBS analysis of geomaterials.(2) Study of sample classification method and development of classificationsoftwareStudy of sample classification method and development ofclassification softwareprovides methodological andsoftware support. PLS-DA classification software has been developed using C#based on.NET framework.SVM classification software hasbeen developed using mixed programming of C#and MATLAB based on type-safeinterface technology.Classification study can be carried out using the data fromgeological sample database since PLS-DA and SVM classification module anddatabase module are integrated in a LIBS desktop application. The developed sampleclassification software meets the design requirements, which can be used for LIBSclassification of geomaterials.(3) Application study in geological field of LIBS database combined with sampleclassification methodThe developed LIBS database for geomaterial combined with sample classificationmethod has been used for application study in geological field.The strong featurepeaks of Si, Al, Ca elements in andesite and carbonate, and the strong feature peaks ofCa, Al, K, Fe in chestnut soil and yellow brown earth are identified using featurespectral database.Lithology identification and soils classification are studied usinggeological database combined with PLS-DA and SVM.The correct classification rateof geomaterial achieves greater than90%using PLS-DA and95%using SVM.Allresults present the combination of LIBS database for geomaterial and classificationmethod can provide a potential tool for geology study based on LIBS technique.