全生育期玉米叶面积指数反演的高光谱智能观测模式研究

The trends of earth observing based on hyperspectral remote sensing technology in future is “intelligent hyperspectral remote sensing satellite system”. Currently, the researches about intelligent hyperspectral remote sensing, which have been focused on how to set up the observation indicators of sensors when toward typical surface scenarios，and already made some progress. However, in the whole growth period of crops, how the spectral indicators and SNR (Signal-to-Noise Ratio) indicators of hyperspectral sensors to be changed, along with the application requirements of parameters retrieve in the different growth period, which was still inadequate. Therefore, the relevant basic researches need to be carried out urgently. In this study, take LAI (Leaf Area Index) of maize for example, the mechanism research of LAI retrieve will be firstly studied in the whole growth period; And then from the point of spectral resolution, center wavelength, band width, SNR and other indicators, the sensitivities of LAI inversion models to spectral indicators and SNR indicators will be discussed in the different growth period. Meantime, quantitative evaluating the influence of observation indicators on inversion accuracy will have to proceed simultaneously; At last, to comprehensive analysis the choice principles and methods of intelligent observation indicators based on LAI inversion accuracies requirements, aim to construct hyperspectral remote sensing intelligent observation model of LAI inversion in a whole growth period of maize. This study will provide the technical supports for hardware adjustment and optimization in hyperspectral intelligent observation mode, and accommodating the future development of intelligent hyperspectral remote sensing. At the same time, the research can be considered as the beneficial supplement for hyperspectral intelligent observation model research under the typical surface scenarios.

高光谱遥感卫星系统的智能化是未来高光谱遥感技术对地观测的发展趋势。目前智能高光谱遥感研究在面向典型地表场景如何设置传感器的观测指标等方面取得一定进展。然而，面向作物全生育期，高光谱传感器光谱指标和信噪比指标应如何伴随生育期参数反演的应用需求而发生改变，此方面研究尚显不足，因此迫切需要开展相关基础研究。本研究以玉米LAI参数为例，开展全生育期玉米LAI参数反演机理研究；从光谱分辨率、中心波长、波段宽度、信噪比等指标入手，探讨不同生育期内LAI反演模型对光谱指标和信噪比指标的敏感性，并定量化评价观测指标对反演精度的影响；根据LAI反演的精度要求，综合分析智能观测指标的取舍原则和方法，构建面向全生育期玉米LAI反演的高光谱智能观测模式。本研究将为高光谱智能观测模式下的硬件调整和优化，适应未来智能高光谱遥感的发展提供技术支撑，同时可为目前正在开展的典型地表场景的高光谱智能观测模式研究提供有益补充。

针对农作物场景对传感器参数指标的需求，开展高光谱传感器的观测指标设置，有效配置遥感器的核心观测指标（中心波长、波段数量、信噪比等）并进行优化选取，有针对性地获取更高效的高光谱数据，进行作物理化参量信息的高精度估算，最终探寻满足全生育期内作物理化参量反演需求的高光谱智能观测模式，可为未来智能高光谱遥感系统发展提供重要的先期实验。目前智能高光谱遥感研究在面向典型地表场景如何设置传感器的观测指标等方面已取得一定进展。然而，面向作物的不同生育期，高光谱传感器光谱指标和信噪比指标应如何伴随生育期参数反演的应用需求而发生改变，此方面研究尚显不足，因此迫切需要开展相关基础研究。本项目以冬小麦和玉米为研究对象，综合利用地面实测成像光谱数据、模拟光谱数据、配套农作物LAI和叶绿素含量等观测数据，通过分析不同生育期内冬小麦和玉米冠层光谱特征，以及LAI及叶绿素含量等高光谱遥感反演机理，研究了高光谱对地观测遥感器的核心观测指标（光谱指标和信噪比指标）对选取作物LAI和叶绿素含量反演的敏感性和有效性；定量化评价了观测指标对LAI和叶绿素含量反演精度的影响，并探寻满足选取作物LAI和叶绿素含量高精度反演需求的有效光谱指标和信噪比指标；最终构建了面向不同生育期内冬小麦和春玉米LAI和叶绿素含量反演的高光谱智能观测模式。本研究将为高光谱智能观测模式下的硬件调整和优化，适应未来智能高光谱遥感的发展提供技术支撑，同时可为目前正在开展的典型地表场景的高光谱智能观测模式研究提供有益补充。