Warren Russell E, Vanderbeek Richard G, Ben-David Avishai, Ahl Jeffrey L
EO-Stat Inc., 10010 Vail Drive, Chapel Hill, North Carolina 27517-7400, USA.
Appl Opt. 2008 Aug 20;47(24):4309-20. doi: 10.1364/ao.47.004309.
We present a sequential algorithm for estimating both concentration dependence on range and time and backscatter coefficient spectral dependence of optically thin localized atmospheric aerosols using data from rapidly tuned lidar. The range dependence of the aerosol is modeled as an expansion of the concentration in an orthonormal basis set whose coefficients carry the time dependence. Two estimators are run in parallel: a Kalman filter for the concentration range and time dependence and a maximum-likelihood estimator for the aerosol backscatter wavelength and time dependence. These two estimators exchange information continuously over the data-processing stream. The state model parameters of the Kalman filter are also estimated sequentially together with the concentration and backscatter. Lidar data collected prior to the aerosol release are used to estimate the ambient lidar return. The approach is illustrated on atmospheric backscatter long-wave infrared (CO2) lidar data.
我们提出了一种顺序算法，利用快速调谐激光雷达的数据，估计光学薄局部大气气溶胶的浓度对距离和时间的依赖性以及后向散射系数光谱依赖性。气溶胶的距离依赖性被建模为浓度在正交基集中的展开，其系数携带时间依赖性。两个估计器并行运行：一个用于浓度距离和时间依赖性的卡尔曼滤波器，以及一个用于气溶胶后向散射波长和时间依赖性的最大似然估计器。这两个估计器在数据处理流中持续交换信息。卡尔曼滤波器的状态模型参数也与浓度和后向散射一起顺序估计。在气溶胶释放之前收集的激光雷达数据用于估计环境激光雷达回波。该方法在大气后向散射长波红外（CO2）激光雷达数据上得到了说明。
