2194 光谱学与光谱分析 第29卷 采样法的有效融合，形成新的减量精细采样法，采用矩阵的 3结论和讨论 形式，在自编Matlab程序下分析发现新的敏感波段，从而实 现了对高光谱信息的快速分析与特征提取。本方法全面系 (1)本文采用减量精细采样法，系统组建了所有两两波 统、简单易行，为高光谱的精细分析及波段探索提供了一种 段形成的RSI指数矩阵，发现了估算小麦叶片氨积累量的最 快捷、方便、科学的研究手段。 佳新高光谱特征波段(990,720)，创建了可靠适用的定量监 (3)对于小麦叶片的氮素监测，不同的植被指数可能有 测模型y=5095‘RSI(990,720)-6040,从而为便携式小麦 不同的监测范围，本研究中RVI(990,720)适用的小麦叶片 氮素监测仪的研制开发及空间遥感信息的分析提取提供了适 氮积累量范围为015~982gN·m2.今后，有必要采用 用可行的核心波段选择。 不同生态点的试验资料对所构建的检测模型进行更为广泛的 (2)本研究针对海量高光谱数据，通过降采样法和精细 测试与评价 参 考文献 【1】Stone ML,Soile」B,Raun W R.Transactions of the ASAE,1996,39:1623 [2 Hansen PM.SchioerringJ K.Remote Sensing of Environment 2003,:542. 【3】Zhang]H,wK,Bailey JS.etal.Pe由osphere,20o6.16(1):108. [4]Tarpley L.Reddy K R.SassenrathrCole G F.Crop Science.000.(6):1814. [5]Gupta R K.Vijayan D.Prasad T S.Advance in Space Research.2003.32(11):2217. 16 学与光请分析)，2006,26(8) 145 光话分析)，2008 [9 Miller J R.Hare E W.WuJ.International Journal of Remote Sensing,1990,11(10):1755. [O1 Blackmer T M Schepers I Varvel G E et al Agronomy lournal 1996 88()-1 [11]Pefuelas Filella I.Biel C.et al.Internal Journal of Remote Sensing,1993,14(10):1887. [12] Zarco-tejada PJ,Miller J R.Noland,et al.IEEE Transactions on Geoscience and Remote Sensing,2001,39(7)1491 13 Xue L H.Cao WX.Luo W H et al.Agronomy Journal,004,9:135. 14 Feng W,Yao x hu Y,et al.European Journal of Agronomy,200 UHr,Zn,TNon,a天华民，朱地阳水起等.AaA0mana作物学报，207.8面 (16]Zhu Y.Li YX.Feng W.etal.Canadian Journal of Plant Science.:1037 Exploring Novel Hyperspectral Band and Key Index for Leaf Nitrogen Accumulation in Wheat YAO Xi,ZHU Yan,FENG Wei,TIAN Yongchao,CAO Werxing' Jiangsu Key Laboratory for Information Agriculture,Nanjing Agricultural University,Nanjing 210095,China Abstract The objectives of the present study were to explore new sensitive spectral bands and ratio spectral indices based on precise analysis of ground-based hyperspectral information,and then develop regression model for estimating leaf N accumulation per unit soil area (LNA)in winter wheat Triticum aestivum L.)Three field experiments were conducted with different rates and cultivar types in three consecutive growing seasons.and time-course measurements were taken on canopy hyperspectral reflectance and LNA under the various treatments.By adopting the method of reduced precise sampling,the detailed ratio spec tral indices (RSI)within the range of 350-2 500 nm were constructed,and the quantitative relationships between LNA (gN m)and RSI (.)were analyzed.It was found that several key spectral bands and spectral indices were suitable for estimating LNA in wheat,and the spectral parameter RSI(990,720)was the most reliable indicator for LNA in wheat.The regression model based on the best RSI was formulated asy=5.095x-6040,with of 814.From testing of the derived equations with independent experiment data.the model on RSI (990.720)had R of 0.847 and RRMSE of 24 7%Thus it is conclude that the present hyperspectral parameter of RSI(990,720)and derived regression model can be reliably used for estimating LNA in winter wheat.These results provide the feasible key bands and technical basis for developing the portable instrument of moni- 994-2010 China Academ Joumal Electronie Publishing House.All rights reserved bttp://www.cnki.ne 3 结论和讨论 (1)本文采用减量精细采样法 , 系统组建了所有两两波 段形成的 RSI 指数矩阵 , 发现了估算小麦叶片氮积累量的最 佳新高光谱特征波段 (990 , 720) , 创建了可靠适用的定量监 测模型 y = 51095 3 RSI(990 , 720)261040 , 从而为便携式小麦 氮素监测仪的研制开发及空间遥感信息的分析提取提供了适 用可行的核心波段选择。 (2)本研究针对海量高光谱数据 , 通过降采样法和精细 采样法的有效融合 , 形成新的减量精细采样法 , 采用矩阵的 形式 , 在自编 Matlab 程序下分析发现新的敏感波段 , 从而实 现了对高光谱信息的快速分析与特征提取。本方法全面系 统、简单易行 , 为高光谱的精细分析及波段探索提供了一种 快捷、方便、科学的研究手段。 (3) 对于小麦叶片的氮素监测 , 不同的植被指数可能有 不同的监测范围。本研究中 RVI(990 , 720) 适用的小麦叶片 氮积累量范围为 0115～9182 gN ·m - 2 。今后 , 有必要采用 不同生态点的试验资料对所构建的检测模型进行更为广泛的 测试与评价。 参 考 文 献 [ 1 ] Stone M L , Soile J B , Raun W R. Transactions of t he ASAE , 1996 , 39 : 1623. [ 2 ] Hansen P M , Schjoerring J K. Remote Sensing of Environment , 2003 , 86 : 542. [ 3 ] Zhang J H , W K , Bailey J S , et al. Pedosphere , 2006 , 16 (1) : 108. [ 4 ] Tarpley L , Reddy K R , Sassenrat h2Cole G F. Crop Science , 2000 , 40 (6) : 1814. [ 5 ] Gupta R K , Vijayan D , Prasad T S. Advance in Space Research , 2003 , 32 (11) : 2217. [ 6 ] Demetriades2shah T H , Steven M D , Clark J A , Remote Sensing of Environment , 1990 , 33 (1) : 55. [ 7 ] LIU Yan2de , YIN G Yi2bin (刘燕德 , 应义斌) . Spectroscopy and Spectral Analysis(光谱学与光谱分析) , 2006 , 26 (8) : 1454. [ 8 ] WAN G Yuan , HUAN GJing2feng , WAN G Fu2ming , et al (王 渊 , 黄敬峰 , 王福民 , 等) . Spectroscopy and Spectral Analysis(光谱学与 光谱分析) , 2008 , 28 (2) : 273. [ 9 ] Miller J R , Hare E W , Wu J. International Journal of Remote Sensing , 1990 , 11 (10) : 1755. [ 10 ] Blackmer T M , Schepers J S , Varvel G E , et al. Agronomy Journal , 1996 , 88 (1) : 1. [ 11 ] Peňuelas J , Filella I , Biel C , et al. Internal Journal of Remote Sensing , 1993 , 14 (10) : 1887. [ 12 ] Zarco2tejada P J , Miller J R , Noland , et al. IEEE Transactions on Geoscience and Remote Sensing , 2001 , 39 (7) : 1491. [ 13 ] Xue L H , Cao W X , Luo W H , et al. Agronomy Journal , 2004 , 96 : 135. [ 14 ] Feng W , Yao X , Zhu Y , et al. European Journal of Agronomy , 2008 , 28 : 394. [15 ] WU Hua2bing , ZHU Yan , TIAN Yong2chao , et al (吴华兵 , 朱 艳 , 田永超 , 等) . Acta Agronomica Sinica (作物学报) , 2007 , 33 (3) : 518. [ 16 ] Zhu Y , Li Y X , Feng W , et al. Canadian Journal of Plant Science , 2006 , 86 : 1037. Exploring Novel Hyperspectral Band and Key Index for Leaf Nitrogen Accumulation in Wheat YAO Xia , ZHU Yan , FEN G Wei , TIAN Yong2chao , CAO Wei2xing 3 Jiangsu Key Laboratory for Information Agriculture , Nanjing Agricultural University , Nanjing 210095 , China Abstract The objectives of the present study were to explore new sensitive spectral bands and ratio spectral indices based on precise analysis of ground2based hyperspectral information , and then develop regression model for estimating leaf N accumulation per unit soil area (LNA) in winter wheat ( T riticum aesti vum L. ) . Three field experiments were conducted with different N rates and cultivar types in three consecutive growing seasons , and time2course measurements were taken on canopy hyperspectral reflectance and LNA under the various treatments. By adopting the method of reduced precise sampling , the detailed ratio spec2 tral indices (RSI) within the range of 35022 500 nm were constructed , and the quantitative relationships between LNA (gN · m - 2 ) and RSI (i , j) were analyzed. It was found that several key spectral bands and spectral indices were suitable for estimating LNA in wheat , and the spectral parameter RSI (990 , 720) was the most reliable indicator for LNA in wheat. The regression model based on the best RSI was formulated as y = 51095 x - 61040 , with R 2 of 01814. From testing of the derived equations with independent experiment data , the model on RSI (990 , 720) had R 2 of 01847 and RRMSE of 2417 %. Thus , it is concluded that the present hyperspectral parameter of RSI (990 , 720) and derived regression model can be reliably used for estimating LNA in winter wheat. These results provide the feasible key bands and technical basis for developing the portable instrument of moni2 2194 光谱学与光谱分析 第 29 卷