基于能量重分配的波长偏移校正方法

针对原子光谱分析中波长偏移问题，提出了一种基于能量重分配原理实现波长偏移校正的方法.用抛物线插值法对能量密度分布进行反演，以谱图相似度为判断指标，通过三分法确定最佳偏移量，并根据最佳偏移量与能量密度分布对光谱强度进行重新计算，实现了亚像素级别的波长校正.对微型光纤光谱仪与单道扫描光谱仪获得的谱图进行了算法验证，使Eu元素的背景噪音平均降低了48%，Mo元素的信号稳定度由20%改善至约5%.该方法对具有结构化特征的谱图具有较好的适用性，且无须寻峰算法配合使用，有助于提高光谱分析精确度.

Wavelength Drift Correction Method Based on Energy Redistribution

In the field of atomic spectral analysis, wavelength stability has a great influence on the results of qualitative and quantitative analysis. In order to solve the problem of wavelength drift, an algorithm for implementing wavelength drift correction based on energy redistribution principle was proposed. The energy density distribution was estimated with a modified parabolic interpolation method, the optimal offset was determined by making use of the similarity of the spectra as the indicator, and then the spectral intensity was recalculated according to the optimal offset and energy density distribution, finally sub-pixel level wavelength drift correction was achieved. The proposed algorithm was verified with spectral data obtained from fiber optic spectrometer and sequential scanning spectrometer. As a result, the background noise of Eu is reduced to about 48% and the signal stability of Mo is improved from 20% to 5%. This algorithm is very suitable for the spectra with structured features, and it will help to improve the accuracy of atomic spectral analysis without peak-searching algorithm.