导数法广泛应用于重叠峰分辨,但信噪比会随着导数阶数的增加而降低。提出一种基于卡尔曼滤波的导数方法,可以显著提高高阶导数谱的信噪比,提高重叠峰分辨力。另外,该方法不需要建立数学模型,计算简便,能满足手持式仪器实时性要求。用仿真数据研究了该法在重叠峰分辨中对信噪比的影响,以离子迁移率谱图分辨为例说明了该法的实际应用。
Derivative method has been widely used for resolving overlapped spectra, but the noise level increases apparently in higher order derivative calculation. A new method is proposed to resolve overlapped spectra based on Kalman filter,which can significantly enhance the signal-to-noise ratio(SNR) at higher order derivative calculation and improve the resolution for overlapped spectra. Moreover,it does not need to develop new mathematical model and its simplicity makes it a plausible real-time solution for portable instruments. The effcets on the SNR for overlapped spectra in practical usage of this method are illustrated using simulated data and IMS spectra.