Abstract：With the multi-resolution analysis, features of infrared spectra of polluted gases were extracted. Then the data were trained or identified by a neural network system. The experimental results show that the combination of the wavelet transform and the neural network has the great ability of feature extracting. And the system is quite efficient for identifying infrared spectra.
刘美娟，冯巍巍，史丰荣，王学勤，张骏*. 污染气体红外光谱特征的快速提取与识别[J]. 光谱学与光谱分析, 2006, 26(10): 1854-1857.
LIU Mei-juan, FENG Wei-wei, SHI Feng-rong, WANG Xue-qin, ZHANG Jun* . Fast Algorithm for Feature Extraction and Identification of Infrared Spectra of Polluted Gases . SPECTROSCOPY AND SPECTRAL ANALYSIS, 2006, 26(10): 1854-1857.
[1] Hoffland L D, Piffath R J, Bouck J B. Opt. Eng., 1985, 24(6): 982. [2] ZHANG Jun, XUN Yu-long(张 骏, 荀毓龙). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 1998, 18(6): 649. [3] Mayfield H T, Eastwood D, Burggraf L W. Chemical and Biological Sensing, 2000, 4036: 54. [4] Tanabe K, Tamura T, Uesaka H. Applied Spectroscopy, 1992, 46: 807. [5] NIE Liang, ZHANG Jun(聂 亮, 张 骏). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2004, 24(8): 953. [6] Mallat S G. IEEE Trans. on Pattern Analysis Machine Intelligence, 1989, 11(7): 674. [7] Daubechies I. Ten lecture on Wavelets. Philadelphies: Capital City Press, 1992. [8] YUAN Zeng-ren(袁曾任). Artificial Neural Network and Its Application(人工神经元网络及其应用). Beijing: Tsinghua University Press(北京：清华大学出版社)，1999. 66. [9] ZHANG Jun, et al(张 骏, 等). J. of Infrared and Millimeter Waves(红外与毫米波学报), 1997, 16(6): 463.
