玉米叶绿素高光谱遥感估算模型研究  

Hyperspectral Remote Sensing Estimation Models for Chlorophyll Concentration in Corn

作　　者：易秋香[1] 黄敬峰[1] 王秀珍 钱翌[3] 

机构地区：[1]浙江大学农业遥感与信息技术研究所,杭州310029 [2]浙江气象科学研究所,杭州310004 [3]新疆农业大学资源与环境学院,乌鲁木齐830052

出　　处：《科技通报》2007年第1期83-87,105,共6页Bulletin of Science and Technology

基　　金：国家自然科学基金项目(40271078和30371018);科技部科技基础性工作专项(2003DEA2C010-13)

摘　　要：对不同品种玉米测定了不同层位叶片(上、中、下)以及叶片不同部位(叶尖、叶中、叶基和叶脉)的室内光谱反射率及其对应的叶绿素含量,采用单变量线性与非线性拟合分析技术,对叶绿素含量与原始光谱反射率、高光谱特征参数(如红边波长、红边位置以及红边面积等)以及植被指数之间的关系进行了分析。结果表明:叶绿素含量与原始光谱在713 nm处具有最大相关系数(R=-0.815),并且基于此波长所构建的指数估算模型明显优于线性模型;基于光谱位置和光谱面积的变量与叶绿素都呈负相关,植被指数与叶绿素含量均呈正相关,并且植被指数相关性优于单一的特征变量,比值植被指数(RVI)相关性优于归一化植被指数(NDVI);模型精度检验结果表明以蓝边面积变量(SDb)为自变量所构建的指数模型对叶绿素含量估算较好。The leaves＇ hyperspectral reflectance and corresponding chlorophyll concentration of six type of corn were measured. The correlations between raw hyperspectral reflectance, hyperspectral characteristic variables, RVI, NDVI and chlorophyll concentration were analyzed. The results indicated that the raw spectral reflectance has the maximum negative correlation coefficient at 713nm and the exponent estimation model was better than linear estimation model. At the meantime the variables based on the spectral areas and wavelengths＇ positions were negatively correlated with chlorophyll concentration, and ratio vegetation indices and normalized difference vegetation indices were positively correlated with it. Chlorophyll concentration in corn leaves was best predicted with the exponential model based on SDb.

关 键 词：玉米 高光谱变量 植被指数 叶绿素含量 

分 类 号：S127[农业科学—农业基础科学]