化学谱图重叠峰检测及基线校正

Overlapping Peak Detection and Baseline Correction of Chemograms

本文提出了一种谱峰曲线与其一阶差分曲线相结合的算法，经宽松变换处理后，用于区分化学谱图中谱峰的终点是否真正落回到基线上，并由此建立了一种具有健壮性的重叠峰检测方法和及时准确的谱图基线校正方法。由于此算法的阈值参数可以固定，从而保证了重叠峰检测较少地受到人为的影响，进一步提高了化学谱图数据处理的自动化水平。

An algorithm which makes use of the chemograms original curve together with its first difference curve,is proposed in this paper.Transformed to be noise tolerant,this algorithm can tell whether a peak end really falls on the baseline.From which ,a robust overlapping peak detection approach and a smart baseline correction approach are derived.With fixed judgement thresholds,the overlapping peak detection and baseline correction are much less influenced by manual factor than before.