基于导数光谱位置变量的干叶片生化组分反演  

Dry-leaf Biochemistry Retrieval by the Position Variables of Derivative Spectra

作　　者：沈艳[1] 牛铮[2] 王汶[3] 徐永明[2] 

机构地区：[1]南京信息工程大学气象学院,南京210044 [2]中国科学院遥感应用研究所,北京100101 [3]中国人民大学环境学院,北京1000872

出　　处：《遥感信息》2005年第4期7-9,34,共4页Remote Sensing Information

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)资助

摘　　要：利用实测干叶片生化组分和高光谱反射率数据,尝试用高光谱位置变量的一阶导数极值特征参数法提取生化组分。结果表明:对全氮、纤维素、木质素、淀粉含量贡献最大的波段分别是D4(1670nm^1735nm反射率一阶导数最小值)、D10(2235nm^2315nm反射率一阶导数最小值)、D8(2020nm^2100nm反射率一阶导数最小值)和D2(1400nm^1480nm反射率一阶导数最小值)。该方法提取干叶片各组分的精度较高,精度最高的是用3个波段提取淀粉含量。分析土壤叶片线性混合光谱表明:该方法能有效的消除土壤背景影响。Using the dry-leaf biochemistry and hyperspectral reflectance data, leaf biochemistry is retrieved through the 1st derivative extremum variables based on hyperspectral position. Results indicate that the highest contributing band on total nitrogen, cellulose,lignin and starch content is D4 （minimum of 1st reflectance derivative at 1 670nm-1 735nm）, D10 （minimum of 1st reflectance derivative at 2 235nm-2 315nm）,D8（minimum of 1st reflectance derivative at 2 020nm-2 100nm）and D2（minimum of 1st reflectance derivative at 1 400nm-1 480nm）, respectively. The precision of dry-leaf biochemistry retrieval is great with highest precision obtained when 3 bands are employed to extract the leaf starch content. The analysis of soil-leaf linear mixed spectra suggests that this method has the potential to effectively remove the soil background effect.

关 键 词：高光谱 位置变量 生化组分 混合光谱 

分 类 号：P237.9[天文地球—摄影测量与遥感]