利用高光谱数据快速估算高羊茅牧草光合色素的研究  

Estimation of photosynthetic pigment of Festuca arundinacea using hyper-spectral data

作　　者：钱育蓉[1,2] 杨峰[1] 李建龙[1] 干晓宇[1] 杨齐[1] 王卫源[3] 

机构地区：[1]南京大学生命科学学院,江苏南京210093 [2]新疆大学软件学院,新疆乌鲁木齐830046 [3]新疆军区司令部机要处,新疆乌鲁木齐830042

出　　处：《草业学报》2009年第4期94-102,共9页Acta Prataculturae Sinica

基　　金：国家高技术(863计划)专题项目(2007AA10Z231)资助

摘　　要：光合色素是植物体进行初级生产的重要物质,能够间接反映植被的健康状况与光合能力,同时,高光谱遥感为快速、大面积监测植被的色素变化提供了可能。本研究以高羊茅牧草为材料,实测了高羊茅冠层的高光谱反射率与光合色素含量数据,对二者进行了相关分析,从5大类12个高光谱特征变量{Rg、R710、Dλr、Dλ700、D[log(1/iλ)]、Sg、RVIa、RVIb、PSSR、PSND、Rch、CARI}中挑选了光合色素敏感参数,建立了植被指数光合色素估算模型。结果表明,光合色素叶绿素a、叶绿素b、总叶绿素和类胡萝卜素与原始光谱进行相关性比较分析时,叶绿素b效果最好;原始光谱、光谱一阶导和倒数对数一阶导3种光谱形式与光合色素进行比较分析时,光谱一阶导在700 nm附近与光合色素相关性最好,相关系数为-0.897;通过12个高光谱特征变量与光合色素相关性比较,选择达到极显著水平的6个变量:Rg、D7λ00、D[log(1/R730)]、RVIb、Rch、CARI进行光合色素含量的回归模型建立,这为利用物理方法快速、无损伤探测高羊茅牧草的营养状况及牧草质量提供了理论支持。Photosynthetic pigments of vegetation are a primary product and important materials that can indirectly reflect health status and photosynthetic capacity of vegetation. Hyper-spectral remote sensing provides the possibility for rapid, large-scale monitoring of vegetation chlorophyll change. In this study, canopy reflecthe possibility for rapid, large-scale monitoring of vegetation chlorophyll change. In this study, canopy reflectance spectra and chlorophyll content of cool-season Festuca arundinacea were measured and their relationship analyzed. The top-five chlorophyll sensitive parameters were selected from 12 hyperspectral characteristic variables, and then used to establish a vegetation index model to estimate photosynthetic pigments. 1）The relationship between Chlb and the original spectrum was the best among the photosynthetic pigments： Chla, Chlb, Chls, and Cars. 2） The first derivative spectra gave a better relationship （correlation coefficient. -0. 897）with vegetation photosynthetic pigments near 700 nm wavelength than the other two hyperspectrum forms {R and DElog（1/R）-]}. Finally, after comparing the correlation between photosynthetic pigment concentrations and hyperspectral data, the most significant variables {Rg, Dλ700, D[log（1/R730）], RVIb, Rch, CARI} were selected from twelve variables to establish the regression model of photosynthetic pigments. All of these provide a theoretical basis for rapid, non-invasive detection of nutritional status and meadow quality of F. arundinacea.

关 键 词：高羊茅牧草 光合色素 高光谱数据 高光谱特征变量 回归模型 

分 类 号：S688.4[农业科学—观赏园艺] S543.9[农业科学—园艺学]