By Ftir Spectra In The Atmospheric Monitoring Applications

Open Path Fourier Transform Infrared (OP-FTIR) Spectroscopy, combined with Computed Tomography (CT), can be used to reconstruct the concentration map of leaking or emitting gas. In the process of real detection, the obtained spectrum always contains some noise in different scale. And the collected data of spectrum rises with the increasing of instruments' precision. In this paper, denoising and compression of spectrum was studied. And the following spectrum after denoising and compressing was reconstructed by the CT algorism, SBFM in this paper. The last section of this paper is the analysis of multi-compounds' spectrum. The following are the major content of this paper.(1) Wavelet analysis was applied to denoise the OP-FTIR spectrum, focused on the study of threshold. False Discovery Rate (FDR) is a method of thresholding which first was studied in this paper. And at that same time, an experienced method for threshold was founded. And the results showed that the two methods are effective for threshold..(2) In this paper, wavelet analysis was also applied to compress the OP-FTIR spectrum. Results showed that this method can compress OP-FTOR spectrums effectively with little deviation.(3) In this paper, the randomly searching derivative method and the first guess method were presented to improve SBFM (the Smooth Basis Function Minimization algorithm). The results indicated that these two cooperative methods were confirmed to be rapid and effective.(4) For the purpose of quantifying each component's content in the mixture, Linear Neural Network was implemented. The results' consistency testified the efficiency of Linear Neural Network in quantitatively analyzing mixture OP-FTIR spectrum.