快速傅立叶变换用于色谱噪声平滑及微弱信号的检测

应用快速傅立叶变换法（ＦＦＴ）对色谱噪声进行平滑处理及微弱信号检测，同时，也与其他数字滤波法进行了比较。结果表明，利用ＦＦＴ法可以很好地对噪声进行平滑处理，使信噪比提高１８倍，为更好地进行痕量组分的色谱微弱信号检测打下了基础。

Application of Fast Fourier Transform Method for Noise Smoothing and Determination of Weak Chromatographic Signal

The primary output of any experiment includes noise. Of fundamental importance to the experimenter is the removal of as much of this noise as possible without, at the same time, unduly degrading the underlying information. Numerical data handling techniques are extensively applied in chromatographic data systems today to enhance effectively the information content, signal to noise ratio( S/N ), accuracy, and precision of analytical signals. In this paper, the principle of the fast Fourier transform (FFT) method has been given. Based on this, the software was developed to smooth noise and improve the signal to noise ratio(SNR) in the trace analysis with high resolution gas chromatography. The proposed FFT smoothing procedure allows to improve SNR by as much as 18 times. In the meantime, the comparison has been made between FFT and other smoothing methods, for example, the moving average method, the least squares smoothing with a seven or five point polynomial and the damped least squares smoothing method.