花生叶绿素含量的高光谱遥感估算模型研究

发布时间：2018-05-18 01:28

本文选题：花生 + 叶绿素含量　；参考：《江苏农业科学》2017年01期

【摘要】：叶绿素是植物体进行光合作用吸收光能物质的主要色素,直接影响植被的光合作用。高光谱遥感为快速、大面积监测植被的叶绿素变化提供了可能。实测了不同品种、肥水条件下,花生冠层的高光谱反射率与叶绿素含量数据,对二者进行了相关分析;首先采用相关系数较大的波段作为变量进行叶绿素含量的估算,其次采用特定叶绿素敏感波段建立叶绿素估算模型。经对比发现,以原始高光谱反射率所构建的估算模型精度不高;一阶导数与叶绿素含量之间的关系采取同样的方法,表明线性模型可较好地预测叶绿素含量;最后在高光谱特征变量中,λr、λg、λo为自变量所构建的模型均通过极显著检验,以λr所构建的指数模型具有最大的决定系数(r2=0.543 5)和F值(F=33.333);通过精度检验,综合分析认为,以662 nm处的一阶微分反射率所构建的线性模型和以红边位置所构建的指数模型均可作为叶绿素含量估算较为合适的高光谱模型。
[Abstract]:Chlorophyll is the main pigment of photosynthesis, which directly affects the photosynthesis of vegetation. Hyperspectral remote sensing provides the possibility for rapid and large area monitoring of chlorophyll changes in vegetation. The hyperspectral reflectance and chlorophyll content data of peanut canopy were measured and analyzed under different varieties and water conditions. Firstly, the band with high correlation coefficient was used as a variable to estimate chlorophyll content. Secondly, a specific chlorophyll sensitive band is used to establish a chlorophyll estimation model. The comparison shows that the estimation model based on the original hyperspectral reflectance is not accurate, and the relationship between the first derivative and chlorophyll content is the same, which shows that the linear model can predict chlorophyll content well. Finally, among the hyperspectral characteristic variables, the models constructed by 位 _ r, 位 _ g and 位 _ o as independent variables have passed the extremely significant test, and the exponential model constructed by 位 _ r has the largest determinant coefficient (r _ 2n _ (0.543 ~ 5) and F value (F ~ (33) 333C). The linear model based on the first order differential reflectance at 662 nm and the exponential model based on the red edge position can be used as hyperspectral models for estimating chlorophyll content.
【作者单位】：山东省农业可持续发展研究所;中国矿业大学;
【基金】：山东省农业科学院科技创新重点项目(编号:2014CXZ09-2)
【分类号】：S565.2;S127

【相似文献】