The Application Of Wavelet Transform In Laser-induced Breakdown Coal Spectroscopy Analysis

Laser-Induced Breakdown Spectroscopy (LIBS), as a rising atomic emission spectroscopy, has been considered as a potential in-situ technology because of its advantages of rapid analysis, small or even no damage to sample, simultaneous multi-elements detection capability and remote distance without contact. In combustion diagnosis field, LIBS could be used for on-line analysis of coal, fly ash and flue gas, which are meaningful for combustion adjusting, operation optimization and pollutant release controlling.Because of noise and large storage,the analysis of LIBS in coal spectrum is limited ,there are many ways to fix the problem .Between those methods ,wavelet is a effective tool .It's a new method develop from FFT ,get good performance in many area .A method based on the wavelet transform (WT) was developed for the compression of laser-induced breakdown coal spectral data. By studing the impacts of main parameters such as the order of db wavelet, decompose level and threshold method on compression performance, it can summarize the compression parameter selection rule and select the proper compression scheme. The scheme was evaluated by compression scores and relative deviation of each spectral line between original and reconstructed. By choosing proper parameters for channel 1, channel 2 and channel 5 of LIBS spectrum of coal sample (No.1~ No.8), the restore score RS and compress score CS are respectively 81%~92.11%, 79.02%~92.07%, with the spectral line relative deviation under 5%. It indicates that the storage space is reduced while the main characteristic of original spectrum is maintained. The result shows that this method is very effective.Wavelet transformation was also applied to denoise LIBS coal spectra. There methods are studided. Db,coiflet and bior wavelet function was used to remove possible noise. The result indicated that when bioresearch 2.8 was applied to a relatively smoother spectra.Finally, the whole research works were summarized and the prospects for the further studies of wavelet on LIBS analysis were suggested.