摘要
提出了一种基于小波变换识别去除毛细管电泳中奇异峰的新方法。电泳图预先去噪,再通过选择合适小波分解,奇异峰可很方便地用一级水平细节识别,然后将其投射到时域中的相应位置去除,去除后的区域采用多项式插值代替。该方法不仅消除了奇异峰,而且有效地降低了背景噪声,正常峰的强度仍保持在90％以上,从而提高了信噪比,降低了检测限。
A new method for recognition and removal of spikes in capillary electrophoresis using wavelet transformation was established in this paper. Firstly, the baseline noise of an electropherogram was removed; Then the spikes could be recognized by their first level detail coefficients after wavelet decompositions of the denoised electropherogram using proper wavelet base. After that spikes locations were projected to the corresponding position in the electropherogram. Finally those spike-removed regions were replaced by polynomial interpolated values. This procedure not only removed the spikes but also significantly reduced the baseline noise while over ninety percent of the normal peak signal remained. Consequently, the signal-to-noise ratio was increased and the detection limit was decreased.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2003年第4期471-473,共3页
Computers and Applied Chemistry
基金
山东省青年科学基金(Q99B11)
教育部"高等学校骨干教师资助计划"项目
