面向精准农业的目标超连续反射光谱获取方法研究

Monitoring crop growth status is the foundation of precision agriculture management. Multispectral/hyperspectral Light Detection and Ranging (LiDAR) is the most effective way to monitoring crop growth status. However, due to the low spectral resolution and less spectral band of the current multispectral/hyperspectral LiDAR, the multiple parameters associated with the crop growth status cannot be measured accurately using the current Multispectral/hyperspectral LiDAR. Aiming at these issues, this project studies acquisition method of objects’ supercontinuum spectrum based on supercontinuum laser pulse and optical fiber dispersion. This method only requires single-pixel detector and single-channel digitizer to obtain objects’ laser echo and supercontinuum spectrum, which not only improves the number and resolution of the obtained spectrum, but also reduces the complexity and cost of hyperspectral lidar. In addition, the method can flexibly adjust the target spectral range and resolution according to the measurement requirements. Research contents include that (1) A mathematical model between objects’ supercontinuum spectrum and laser echo is established; (2) An inverse problem model to obtain objects’ supercontinuum spectrum is established; (3) The objects’ supercontinuum spectrum is rapidly, robustly and accurately extracted using regularization and optimization methods; (4) Simulation and experiment were carried out to verify the performance of the proposed method. The research in this project can effectively improve the number of the spectral bands obtained by hyperspectral LiDAR and spectral resolution of hyperspectral LiDAR, reduce the cost of hyperspectral LiDAR, and provide data and verification platform for the measurement of crops’ growth parameters based on the fine spectrum of crops’ leaves.

农作物形态和多种生化参数的准确、低成本监测是现代精准农业的必备前提，高光谱激光雷达是现阶段可同时测量作物形态和生化参数的唯一手段。本项目针对高光谱激光雷达由于光谱波段少、分辨率低、成本高，无法满足精准农业对农作物生长参数监测的要求，研究面向精准农业的目标超连续反射光谱获取方法。首先，建立目标反射光谱与不同时刻目标反射回波强度之间的连续“时-谱”模型，为目标连续光谱的获取提供理论基础；其次，研究基于离散“逆问题”的目标光谱提取方法，实现目标超连续反射光谱的快速、稳定、高分辨率获取。最后，搭建激光雷达实验平台，获取不同生长状态农作物叶片超连续反射光谱，验证本项目提出的目标超连续反射光谱获取方法。本项目研究内容可有效提高高光谱激光雷达的光谱分辨率，降低高光谱激光雷达的成本。同时，为基于叶片精细光谱的农作物、植被生长参数反演研究提供数据支持和验证平台，使高光谱激光雷达更好地服务于精准农业、林业。

农作物形态和多种生化参数的准确、低成本监测是现代精准农业的前提，高光谱激光雷达是现阶段可同时测量作物形态和生化参数的唯一手段。本项目针对现有高光谱激光雷达由于光谱波段少、分辨率低、成本高，无法满足精准农业对农作物生长参数监测的需求，研究面向精准农业的基于超连续谱激光器和单模光纤色散的目标超连续反射光谱获取方法。主要研究内容包括：目标反射回波强度与目标超连续反射光谱之间“时-谱”关系建模；目标超连续反射光谱的快速、稳定、高精度反演方法研究；基于全波形回波的目标距离在线测量方法；超连续谱激光雷达实验平台设计与研制；主要研究结果包括：建立了基于单模光纤材料色散的目标反射光谱与不同时刻目标反射回波强度之间的连续“时-谱”模型，分析了所提方法的光谱分辨率与光纤色散斜率、采样频率之间的关系，提出了连续“时-谱”模型的离散化方法以及基于非线性优化的目标反射光谱反演方法，并针对典型地物光谱数据（方解石、棉花纤维等）对“时-谱”模型和反演方法进行了仿真验证，结果表明所提方法可实现目标反射光谱高分辨率获取，提出了一系列基于全波形回波的目标距离在线测量方法，设计并研制了高光谱激光雷达实验平台，研制了高速光电探测器和用于光学元件集成的工件，实现了实验平台的小型化集成，获取了不同含水量叶片，不同种类矿石近红外波段反射回波。项目执行期间以通讯作者发表SCI论文4篇，培养硕士毕业生2名。本项目研究内容和成果可有效提高高光谱激光雷达的光谱分辨率，降低高光谱激光雷达的成本。为基于三维形态和精细光谱的农作物、植被生长参数反演提供数据支撑和验证平台，使高光谱激光雷达更好地服务于精准农业、林业。此外，本项目提出的目标超连续反射光谱获取方法可进一步推广到矿产探测和开发领域，实现矿产三维形貌和矿物含量的同时获取，为矿产探测与高效开发提供支撑。