基于导数直方图和神经网络的色谱基线提取算法

Extraction of baseline of chromatography based on histogram of first derivative HPLC and neural networks

根据基线信号的变化特征 ,提出了一种新的色谱基线提取算法 -基于一阶导数直方图和神经网络的色谱基线提取算法。鉴于信号总是存在着噪声 ,含噪的基线信号一阶导数值也较大 ,并且分布无规律 ,而采用了多分辩率小波变换滤波技术。通过选择阈值门限 ,本算法可精确提取出信号的基线 ,尤其适用于非基线对称类的色谱信号的基线 (趋势函数 )

A new algorithm was presented to extract the baseline of chromatography. The algorithm is based on neural networks and the histogram of chromatograpy′s first derivative. Then wavelet is used to filter noise. The algorithm can recover the baseline of chromatograpy exactly by selecting a proper threshold. It is useful to recover baselines of signal from heavy noise.