中文摘要：

采用列文伯格-马夸尔特算法（L—M）和最小二乘法（LS）对带有高斯白噪音的两种模拟衰荡信号进行了拟合分析．结果表明：L—M相对于LS所得结果更准确，相对误差更小，抗噪音能力更强，且重复性更好．搭建了两种光腔衰荡光谱法系统，分别得到连续和脉冲衰荡信号，依据理论分析结果对实验数据进行了处理，所得结论与模拟结果一致．

英文摘要：

Searching an efficient algorithm for extracting ring-down time in cavity ring-down spectroscopy （CRDS） is very important. Two kinds of simulated ring-down signals with Gaussian white noises were analyzed using Levenberg-Marquardt （L-M） algorithm and least squares （LS） method,nd the results show that L-M algorithm gives better results, smaller relative error, higher noises immunity and better repetition than LS method. Two kinds of CRDS setups were founded to produce CW and pulsed ring-down signals, experimental data were processed according to theoretical analyses. Experimental results are consistent with the simulated ones.