摘要
高光谱遥感是对地观测的重要技术手段 ,利用野外光谱仪在地面的实测工作为其在精准农业等方面的应用进行有益的尝试 ,可以提高作物营养诊断的精度。养分胁迫下高光谱特征提取是这一目标的地面预研究。本次试验对不同氮素养分胁迫下冬小麦的不同生育期 ,分别观测其光谱反射率 ,分析其生物物理参数的变化规律和反射光谱的特征 ,利用导数光谱技术对叶绿素含量、叶绿素密度和叶面积指数等生物物理参量同原始光谱、一阶导数光谱进行拟合度比较 ,结合养分胁迫的特点 ,分析建立光谱模型的可能性。叶绿素密度在可见光和近红外波段均拟合较好 ,拟合度在 0 5左右 ,高于叶绿素含量与光谱反射率的拟合 ,而叶面积指数与冠层光谱拟合在近红外波段较好 ,在可见光波段拟合度较小。通过作物的光谱特征 ,提取其中重要的近红外反射峰值、绿峰和红端位移特征 ,与冬小麦的叶绿素密度、叶绿素含量等生物物理参数进行相关分析 ,建立了线性光谱模拟模型。提取出的特征参量均可有效地模拟生物物理参数的变化 ,其中叶绿素含量与近红外反射峰值的拟合度 (R2 )除乳熟期外 ,都在 0 9以上 ;孕穗期反射峰值和绿峰差异最为显著 ,红端位移最大 ,同时根据上述模拟模型的分析 ,孕穗期的模拟模型也有显著的相关关系。综合分析 ,孕穗期是利?
Hyperspectral remote sensing is not only an important technical method in the study of global observation and the vegetation cover change, but also a main study aspect of precision agriculture. This paper first discusses sensors characteristics and image mechanisms of hyperspectral remote sensing data, its image processing, information extraction methodology, image analysis and model identification, as well as potential application of airborne and space borne hyperspectral remote sensing data. In order to monitor crop nutrient supply condition and to realize precision fertilization, different N nutrient stress condition is studies. A series of experiment have been arranged in Luancheng eco system experimental station, which concerns N, P, K stress, respectively. Rank difference of nitrogenous nutrition level is obtained by the man make style through 5 years' field experiment of winter wheat and summer corn. Canopy and plant spectral of different nitrogen application is measured by Analytical Spectral Devices (ASD Field Spec HH TM ) in different stage. The study has analyzed the variation of winter wheat canopy reflectance with wavelength and time, compared the reflectance spectra curves of difference growing stage and N stress condition. According to hyperspectral characteristic, reflectance and first order derivative spectra is determined by wheat biophysical parameters and agricultural properties. Three variables of hyperspectral in association of spectral position are used in the analysis. There are peak value of Near Infrared (NIR), green reflected peak value and red edge shift. The estimate models about biophysical parameters of winter wheat have been built on the basis of canopy reflectance data. The best parameter is chlorophyll density in estimated models of the biophysical parameters and best observational stage is pregnant period to winter wheat. The first derivative spectrum is more correlatively with the biophysical parameter including chlorophyll density and LAI.
出处
《资源科学》
CSSCI
CSCD
北大核心
2003年第1期86-93,共8页
Resources Science
基金
中国科学院知识创新工程项目 (KZCX2 - 4 0 4 - 3)
"973"项目专题 (G2 0 0 0 0 7790 5 - 0 1 )
中科院"百人计划"生态系统管理的基础生态学过程研究项目共同资助
关键词
养分胁迫
高光谱
冬小麦
生物物理参数
相关分析
Nutrient stress
Hyperspectral
Winter wheat
Biophysical parameters
Regression analysis
