针对光腔衰荡光谱(CRDS)的吸收光谱曲线,利用MATLAB工具软件和最小二乘算法对常用的6种气体吸收光谱拟合算法(VP、GP、AVP、AGP、SDVP、SDAVP算法)进行编程,并设计了参数设置、拟合结果显示等界面化窗口;通过CRDS采集已知浓度CO2气体的吸收光谱;采用6种算法对吸收光谱进行拟合实验,比较6种不同线型算法的拟合误差。拟合结果显示:拟合曲线和测量点几乎完全重合;AVP算法拟合的平均残差最大,为3.1748×10 -3, GP及AGP算法的平均残差最小,为1.3212×10 -3,由此验证了6种线型算法在气体吸收光谱曲线拟合中的可行性;AVP算法的运算时间最短,为1.7516 s,SDAVP的运算时间最长,为389.682 s,GP算法的运算时间小于AGP算法。综合考虑残差与运算时间后认为,GP算法最适用于CRDS光谱线型的拟合。

For the spectral absorption curve of cavity ring-down spectroscopy (CRDS), this study uses the MATLAB software and the least-squares algorithm to program six different line-shape fitting algorithms (VP, GP, AVP, AGP, SDVP, and SDAVP algorithms) commonly used for gas absorption spectroscopy and designs interface windows for the parameter setting and fitting result display. The absorption spectrum of the carbon dioxide gas with known concentration is obtained using CRDS. The six aforementioned algorithms are used to fit the absorption spectrum, and the fitting residuals of six different line-shape fitting algorithms are compared. The fitting results denote that the fitting curve is almost coincident with the measuring points. The average AVP residual is the largest, which is 3.1748×10 ^-3, whereas GP and AGP denote the minimum average residual of 1.3212×10 ^-3. This study verifies the feasibility of the six aforementioned line-shape fitting algorithms with respect to the fitting of gas absorption spectrum. The AVP's operation time is the shortest(i.e., 1.7516 s), whereas the SDAVP's operation time is the longest (i.e., 389.682 s). The GP's operation time is less than that of AGP. Therefore, the GP is most suitable for performing the CRDS spectral line-shape fitting with respect to the residuals and operation time.