基于米散射-拉曼激光雷达的高精度大气能见度探测和反演方法研究

Atmospheric visibility is an important meteorological parameter. It is of great significance to obtain its information in real time and accurately for transportation support and environmental monitoring and forecasting. The traditional single wavelength Mie scattering lidar measurement method needs to assume or approximate the lidar ratio and the boundary value in the inversion process, which restricts the accuracy of the inversion result. Based on the characteristics of inversion of extinction coefficient without assuming the lidar ratio and the boundary value for Raman lidar, this project proposes a method of accurate measurement of visibility by using Mie-Raman scattering lidar. The main research contents include: First, based on the Mie scattering lidar, the technology of laser wavelength expansion and multi-channel signals acquisition are studied, and a joint Mie-Raman detection platform is constructed. Then, high efficiency adaptive signal denoising is realized by applying the variational mode decomposition (VMD) algorithm to improve signal to noise ratio of the echo signal. Third, the Ansmann method is used to obtain accurate lidar ratio based on multi-channel Mie scattering and Raman echo signals. Next, accurate correction of the boundary value is performed based on the correlation between the extinction coefficient profiles obtained by the Raman and the Mie scattering lidar, and then the inversion accuracy of the Mie scattering lidar is improved. On this basis, the two-coefficient correction model is set up to achieve accurate correction and calibration of visibility with the measurement result of the transmission visibility meter as the standard. This method can not only improve the inversion accuracy of the Mie scattering lidar, but also break through the bottleneck of the detection distance limitation of Raman lidar.

大气能见度是重要的气象参数，实时准确地获取其信息对于交通运输保障、环境监测预报等具有重要意义。传统的单波长米散射激光雷达测量方法在反演过程中需对雷达比和边界值作假设或近似处理，制约了反演结果的精度。本项目基于拉曼激光雷达反演无需假设雷达比和边界值的特点，提出利用米-拉曼散射激光雷达联合精确测量能见度的方法，主要研究内容包括：基于米散射激光雷达，研究激光器波长扩展、多通道信号采集等技术，构建联合探测平台；利用变分模态分解法实现高效地自适应信号去噪，提高回波信号信噪比；基于多路米散射和拉曼回波信号，采用Ansmann法获取准确的雷达比；依据拉曼和米散射消光系数反演结果的相关性精确修正边界值，提高米散射通道消光系数廓线的反演精度；在此基础上，以透射仪测量结果为标准，建立二系数修正模型，实现能见度的精确修正和标定。该方法不仅能提高米散射激光雷达的反演精度，而且可突破拉曼激光雷达探测距离受限的瓶颈。

大气能见度的测量是进行气象分析的一个重要因素，能见度可以反映当时大气的物理光学状态，预示天气的变化，不仅可以为气象观测部门提供重要分析数据，而且在航空、水上交通、陆上交通以及军事活动等领域都有重要影响目前，这些领域对能见度测量的需求也越来越高。根据柯西米德方程可知，计算大气能见度的关键是实现大气消光系数的反演。在项目实施过程中，主要完成了：（1）基频光源的设计与优化，基于多种新型可饱和吸收体在不同类型激光器中的研究，主要包括新型可饱和吸收体的制备与表征及其在激光器中的应用与测试。（2）激光雷达回波信号的噪声滤除方法研究，提出了变分模态分解、结合局部均值分解和改进阈值法等去噪算法，在激光雷达回波信号的噪声处理上取得了良好效果。（3）基于激光雷达数据的行星边界层高度反演算法研究，通过所设计的WCT-WOA-TL自适应算法，得到的大气边界层高度相比传统算法更为准确。（4）基于激光雷达数据的气溶胶特性反演及分类算法研究，通过建立深度信念网络实现气溶胶消光系数反演以及构造朴素贝叶斯分类器模型识别气溶胶类型。（5）基于激光雷达数据的云层区域反演方法研究，通过改进的微分零交叉法和聚类分析进行云层检测以及双向重构微脉冲激光雷达后向散射信号进行云层参数的反演。本项目的研究结果对大气能见度、气溶胶和云的准确探测以及激光雷达系统的应用推广具有十分重要的理论和现实意义。