红外光谱谐波检测系统中的噪声与基线漂移问题一直是光谱处理的热点, 提出一种采用小波变换的Mallet分解算法, 解决谐波检测中各种复杂噪声以及基线漂移的问题。 选取适当小波函数及分解层次将谐波曲线中含有的噪声和基线漂移与有用信号分解到不同频带; 分析频带信息, 设定一个检测信息频带, 应用阈值处理及系数置零的方法使频率处于此频带的信息保留下来。 小波变换方法可以在一次分解与重构过程中同时去除谐波信号的噪声与基线的双重干扰, 从而将谐波信号有效地测量出来。 实验证明, 应用小波变换进行谐波校正的方法可应用于不同的谐波检测系统, 具有普遍适用性。

The problem of noise and baseline drift is a hot topic in infrared spectral harmonic detection system. This paper presents a new algorithm based on wavelet transform Mallet decomposition to solve the problem of eliminating a variety of complex noise and baseline drift in the harmonic detection. In the algorithm, the appropriate wavelet function and decomposition level were selected to decomposed the noise, baseline drift and useful signal in the harmonic curve into different frequency bands. the bands’ information was analysed and a detecting band was set, then the information in useful frequency was reserved by zeroing method of treatment and the coefficient of the threshold. We can just use once transform and reconstruction to remove interference noise and baseline from double-harmonic signal by applying the wavelet transform technique to the harmonic detection spectrum pretreatment. Experiments show that the wavelet transform method can be used to different harmonic detection systems and has universal applicability.