植物叶片多种色素高光谱遥感机理与模型研究

The studies on the quantitative inversion of leaf pigments, including chlorophyll a, chlorophyll b, beta-carotene, xanthophyll and anthocyanins, using hyperspectral remote sensing technology are the basis of vegetation hyperspectral remote sensing because the pigments content are the indicators of photosynthesis, nutritional status and environmental change. Among studies on content estimation of foliar pigment in the literature reviewed, the estimation of total chlorophyll content is highly performed, and only a few articles were found for the estimation of chlorophyll a or chrolophyll b content.To our knowledge, few studies have tried to relate the beta-carotene content, xanthophyll content and anthocyanins content to the leaf hyperspectral parameters.This is because that the precise measurement of chlorophyll a, chlorophyll b, beta-carotene, xanthophyll and anthocyanins is difficult using spectrophotometer and the overlap phenomina in the leaf spectra of the different pigments. The aim of our work include: 1. According to the pigment varieties and content of plant leaves in different growth stage, Loropetalum Chinense, Acer saccharum Marsh, Elaeocarpus Sylvestris are selected as tested plants preliminarily. Leaves with different pigment content are collected as the experimental material in different growth stage. The reflectance spectra of leaves are measured immediately. The content of chlorophyll a, chlorophyll b, beta-carotene, xanthophyll and anthocyanins will be measured precisely using High Performance Liquid Chromatography. Other parameters such as water content, dry matter, cellulose content, lignin content will be measured at the same time. 2. Spectral peak separation technology which is widely used in Chemometrics will be introduced to data processing of leaf spectra in order to separate the influnce of chlorophyll a, chlorophyll b, beta-carotene, xanthophyll and anthocyanins on leaf spectra, to quantify the overlapping and sheltering characteristics of various pigment on the leaf spectra, and to obtain a spectrum which can reflect the characteristics of a individual pigment, and then to establish estimation models of pigment content using neural network with hyperspectral data. 3.The PROSPECT-5 model will be modified by introduce the input parameters including the content of chlorophyll a, the content of chlorophyll b, the content of β-carotene, the content of xanthophylls, and the content of anthocyanins. The parameters of pigment specific absorption coefficients, leaf structure parameter and leaf refractive index will be obtained by iterating based on the minimum cost function in the range of 400nm-800nm using the leaf reflectance spectra and the corresponding pigment content. After sensitivity and accuracy analysis were performed on modified PROSPECT model,the pigment content, including chlorophyll a, chlorophyll b, β-carotene, xanthophylls, and anthocyanins, can be inversed quantitatively.

由于植物叶片中的色素（叶绿素a和b、β-胡萝卜素、叶黄素和花青素）含量与植物的光合作用、营养状况和环境变化有关，因此，开展植物多色素高光谱遥感机理和模型研究，是植被高光谱遥感的基础工作。本研究利用高效液相色谱仪精确测定叶片各种色素含量，针对叶片中色素吸收波段谱带交叠现象，将化学计量学中的光谱分峰技术引入到植物色素高光谱遥感，量化各种色素在叶片光谱上的重叠和遮蔽特征，获得能反映单一色素的光谱，再采用神经网络方法建立叶片多色素含量高光谱遥感估算模型；将PROSPECT-5模型中的叶绿素、类胡萝卜素含量输入项改为叶绿素a和b、β-胡萝卜素、叶黄素和花青素含量，通过最小代价函数迭代确定模型参数，建立能反映叶绿素a和b、β-胡萝卜素、叶黄素和花青素含量变化的PROSPECT模型，并进行模型的敏感性分析和精度验证；最后，利用改进的PROSPECT模型进行叶片多色素定量遥感反演。

①在国际上获得首套细分叶片叶绿素a、叶绿素b、胡萝卜素、花青素含量和对应光谱数据集，为开展植物叶片多色素遥感反演机理研究打下良好基础；同时采集了基于偏振光遥感的叶片光谱数据集、冻害条件下的油菜叶片色素含量动态变化数据集及其对应的光谱。.②使用高斯-洛伦茨（G-L）函数定义叶绿素a、叶绿素b、类胡萝卜素和花青素吸收系数中的每个吸收峰，解决PROSPECT模型不能分离绿素a、叶绿素b类胡萝卜素和花青素吸收系数问题；基于LOPEX_ZJU数据库，率定叶绿素a、叶绿素b、类胡萝卜素类胡萝卜素和花青素吸收系数及叶片折射系数；构建能够反演叶片叶绿素a、叶绿素b、类胡萝卜素和花青素含量的模型PROSPECT-MP+,并与PROSPECT-D比较验证该模型的光谱模拟与色素反演精度。.③叶片镜面反射对色素含量反演的干扰及去除方法研究：主要包括利用叶片偏振观测，估算并去除叶片镜面反射，再建立叶片色素估算模型，提高叶片色素反演精度；比较了辐射传输模型、混合模型、植被指数等专门考虑了叶片镜面反射和镜面效应的方法，并重点研究了小波变换技术。结果表明，利用小波变换可以较好消除叶片镜面反射的影响，提高叶片色素含量的反演精度。.④油菜叶片冻害高光谱响应机理及色素含量监测方法研究：首次系统研究了油菜叶片在正常、过冷却、受冻、融化后的叶片光谱特征，建立了受冻条件下油菜叶片色素含量最佳光谱监测模型。

内容获取失败，请点击重试