基于导数光谱变量叶片含水量模型的建立  

Establishment of Leaf Water Content Models Based on Derivative Spectrum Variables

作　　者：沈艳[1] 牛铮[2] 王汶[2] 徐永明[2] 

机构地区：[1]南京信息工程大学气象学院,江苏南京210044 [2]中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101

出　　处：《地理与地理信息科学》2005年第4期16-19,共4页Geography and Geo-Information Science

基　　金：国家重点基础研究发展规划项目(G2000077902);中国科学院知识创新工程重大项目(KZCX1-SW-01-02);国家自然科学基金资助项目(40271086)

摘　　要：利用实测单子叶、双子叶植物叶片含水量和高光谱反射率数据,基于导数光谱变量分析方法,讨论利用含水量识别两种植物的可行性,并分别建立两种植物叶片含水量的单变量和多变量估算模型。结果表明:D2、D6和D8等含水量表征方法能在一定程度上识别单、双子叶植物;单变量估算模型中,等价水深(EWT)与高光谱参数的相关性更大,单子叶植物的相关性大于相应的双子叶植物,并且以简单的一元线性模型普适意义最大,反演效果最好;多变量模型中,用FMC表征单子叶叶片含水量精度大于EWT,用EWT表征双子叶叶片含水量精度大于FMC,总体反演精度是单子叶植物高于双子叶植物。Based on the measured leaf water content Cw （as below） and hyperspectral reflectance data of monocotyledon and dicotyledon,derivative spectrum variables are used to discuss the feasibility of identifying the monocotyledon and dicotyledon. On the other hand, single and multiple variable estimation models of their leaf Cw are established. Research result indicates that variables such as D2、D6 and D8 have the leaf- type identification ability, Single variable model suggests that the correlation degree between equivalent water thickness （EWT） and hyperspectral parameters is better than corresponding furl moisture content （FMC） and the parameters for different Cw demonstrator, and the corresponding correlation degree of monocotyledon is higher than the dicotyledon＇ s. It shows that simple unitary linear model has the most universal significance and forecast results. Multiple variable model indicates that precision is higher when monocotyledonous Cw is expressed as FMC and dicotyledonous Cw is expressed as EWT. But the whole forecast precision of monocotyledon is greater than dicotyledon＇ s.

关 键 词：单子叶植物 双子叶植物 含水量模型 导数光谱变量 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]