基于地物波谱库构造农作物生长参数的时空分布先验知识

认识典型地物的波谱特征及其随波谱测量环境变量的变化,从积累的观测数据中提取地物波谱及其配套参数的时空分布知识,是对地表参数定量遥感反演研究的重要支撑.基于"中国典型地物标准波谱数据库"中收集的数据和模型,对典型农作物测量数据进行了系统整理,得到了我国5种农作物的主要生长参数(叶面积指数、全氮含量)在其主要生育期的统计数据,构造了这些参数的时空分布知识.文中所述的数据和方法,对于遥感反演中地表参数先验知识的初始值与不确定性估计、数据同化等相关研究,都有使用和参考价值.

THE PRIOR KNOWLEDGE OF CROP GROWTH PARAMETERS CONSTRUCTED FROM THE SPECTRUM DATABASE OF LAND SURFACE OBJECTS

It is an intense problem for both remote sensing fundamental research and its application to investigate the reflectance spectrum feature of typical objects and its variance under different measurement conditions.In this context,extracting the temporal and spatial distribution of the knowledge about spectrum data and their environmental variables from the accumulated measuring data is an important support for quantitative remote sensing modeling and the retrieval on land surface parameters.To achieve above aims,the biophysics and biochemistry parameter data of 5 types of crop collected by the "863" Project are conducted in this paper.The statistical data of major growth parameters of 5 crop species on major growing stages in China,such as LAI,Nitrogen are gotten.Based on the statistic result,the tempo-spatial distribution knowledge is obtained borrowing the relationship between the LAI and NDVI when the parameter of LAI is considered.Thus,this type of knowledge can be used to build the background field of required initial data while inversing a physical model to estimate the land surface parameters.The proposed method to calculate the background values of parameters and to extend these in situ statistical knowledge on a tempo-spatial scale,as well as the results presented in this paper can be considered as valuable reference while estimating the initial value and its associated uncertainty in physical model inversion or data assimilation.