Abstract: In order to eliminate the signal noise of the direct absorption spectrum，the separation of high and low frequency portions of the signal noise was achieved by using wavelet analysis. The direct absorption signal was generated by several steps，such as light generation by distributed feedback (DFB) diode laser，gas absorption，data acquisition. Using the method mentioned above，the high frequency noise of the signal was eliminated. After the de-noising experiment of two signals，the de-noised signals reflected the spectral characteristics of the data，which is propitious to the fitting process of the spectral data. In particular，the giant noise in the strong interfered signal is de-noised，so that the signal can show the speed characteristics of the air flow which can perform the following step of lineshape fitting.

Key words: Wavelet analysis；Direct Absorption Spectroscopy；TDLAS；Noise cancellation；Distributed feedback diode laser

摘要： 采用小波分析方法对直接吸收谱信号进行消噪处理，通过对信号的高、低频部分进行分离实现了噪声消除。采用这种方法对近红外波段的DFB(分布反馈)半导体激光器所发出并且经由气室吸收、数据采集等步骤产生的直接吸收信号进行噪声消除实验，清除了数据中的高频噪声。实验中的两种光谱信号经消噪处理后，反映了数据的光谱特性，有利于进行光谱数据的拟合处理；尤其将受到气流强烈干扰的信号去除了大幅值的噪声信号，使该信号反映气流的速度特性，并能进行后续的线型拟合工作。

关键词: 小波分析；直接吸收谱；可调谐半导体激光吸收谱技术；噪声消除；分布反馈半导体激光器