ASD Field Spec3野外便携式高光谱仪诊断冬小麦氮营养

氮素营养诊断关键在于氮营养指数(nitrogen nutrient index,NNI)预测。对于冬小麦氮营养指数预测模型而言,如何选取预处理方法和建模方法不一而足,不同预处理和模型选取对预测结果精度的影响程度目前还不清楚。该研究以ASD Field Spec3野外便携式高光谱仪采集乐陵市冬小麦冠层高光谱数据,采用10种光谱预处理方法并结合3种模型(偏最小二乘回归、BP神经网络和随机森林算法)建立多种冬小麦氮营养指数高光谱预测模型。对比模型预测精度表明最佳的高光谱建模方法为随机森林算法结合SG卷积平滑预处理所建模型(预测集R2=0.795,RMSE=0.125,RE=11.7%)精度高、可靠性强,是筛选出最佳的冬小麦氮营养指数高光谱预测模型。该研究结果对冬小麦氮营养指数高光谱预测建模具有科学价值,为筛选最优高光谱预处理方法和预测模型提供技术参考。     

基于棉田原位高光谱数据的土壤pH值反演与制图研究

[目的] 研究快速、准确大面积监测农田土壤pH值,为大面积土壤改良和实现农田精细化管理提供科学支持。[方法] 以南疆阿拉尔市十二团棉田为研究区,采用网格采样法采集231个样点的原位高光谱数据,并同步采集其中116个样点的土壤样品;分析了原位高光谱反射率数据经不同预处理模式后的光谱数据与土壤pH值的相关性;采用偏最小二乘回归、支持向量机回归和随机森林3种建模方法分别建立了土壤pH值的高光谱反演模型,根据模型评价指标优选出最优模型对未采集土壤样点的pH值进行反演制图。[结果] 反射率经微分处理后可有效改善其与土壤pH值的相关性;反射率二阶微分的随机森林模型为所有模型中的最优模型,其验证集的R²为0.87,RMSE为0.04,RPD为2.53;最优模型反演的pH值数据插值所得数字图与实测pH值插值图的空间分布特征高度吻合,能客观反映土壤碱化的空间分布状况。[结论] 随机森林模型为原位反演南疆棉田土壤pH值的最优模型,克里金插值能够客观可视化表达研究区土壤pH值的分布状况。     

多种潮土有机质高光谱预测模型的对比分析

为了对比不同方法建模效果的差异,筛选潮土有机质高光谱最佳预测模型,该研究采集国家潮土土壤肥力与肥料效益长期监测站不同施肥处理耕层土样83份,采用25种光谱预处理方法(15种单一预处理方法,10种预处理方法相加算法)结合3种建模方法(多元线性回归、偏最小二乘回归、支持向量机回归),构建不同的潮土有机质高光谱预测模型。对比模型预测结果表明,最佳光谱建模方法是偏最小二乘回归法,该方法结合多种预处理方法均获得了较高的模型预测精度和可靠性,25个检验模型的平均决定系数、均方根误差值RMSEv和相对分析误差RPD值分别为0.913、1.264 g/kg和3.299。使用预处理方法相加算法能更好地提升模型精度,相比使用单一预处理方法,3种建模方法的检验模型平均决定系数分别提高了0.049、0.033和0.071,RMSEv分别降低了0.318、0.204和0.528 g/kg,RPD值分别提高了0.530、0.307和1.144。先用多元散射校正法再进行5个平滑点数的一阶导数预处理在3种建模方法中表现均较好(平均决定系数=0.934,平均RMSEv=1.17 g/kg,平均RPD=3.59),可作为潮土有机质预测模型的通用预处理方法。偏最小二乘回归模型结合最大值标准化预处理所建模型(决定系数=0.948,RMSEv=0.972 g/kg,RPD=4.276)精度高、可靠性强,且建模过程数据运算更为简便,是筛选出的最佳潮土有机质高光谱预测模型。该研究结果对潮土有机质高光谱预测建模有一定的指导作用,并为筛选最佳高光谱预测模型提供技术参考。     

基于高光谱图像的牧草粗蛋白含量反演模型

粗蛋白(crude protein,CP)是评价牧草品质和饲用价值的重要指标。利用高光谱技术实现大面积牧草CP含量实时、准确、无损监测是草地营养状况监测的重要内容。为掌握青海省海晏县金银滩草原牧草CP含量的分布状况,该文采用课题组航空飞艇搭载自主集成高光谱成像系统获取高分辨率高光谱图像,对高光谱数据进行光谱衍生变换,采用不同建模方法构建CP含量的反演模型。选取最佳组合的2种光谱指数为自变量构建牧草CP含量的单变量模型。对于包络线去除的衍生光谱和对数、归一化、一阶微分及其衍生组合变换光谱,使用逐步判别分析法筛选各光谱变量的特征波段作为自变量,构建牧草CP含量的多元线性和非线性回归模型,综合比较各模型的精度选择最优反演模型。结果表明,不同光谱变量相比,微分光谱变量对牧草CP含量拟合效果较好,R~2均达到0.794以上。不同多元回归模型相比,非线性回归模型精度高于对应的线性回归模型。以光谱对数后再一阶微分变量(D(log(R)))构建的多元非线性回归模型为牧草CP含量最优估算模型,R~2为0.918,RMSE为0.054。将D(log(R))建立的非线性回归模型应用于高光谱图像上,绘制研究区牧草CP含量空间分布图。研究为大区域尺度CP含量的定量反演及精准畜牧业的高效实施提供参考和技术依据,也为今后智慧畜牧业的发展奠定基础。     

基于高光谱的复垦农田土壤有机质含量估测

为了快速准确估算矿区复垦土地土壤有机质含量,以永城矿区复垦农田为例,在土样有机质含量测定和高光谱数据测量的基础上,对土壤高光谱数据进行多种预处理并与有机质实测含量进行相关性分析,利用相关系数进行P=0.01水平显著检验,确定敏感波段,建立一元线性回归、多元逐步回归和偏最小二乘回归等多种有机质含量与高光谱估测模型。结果表明:经过数学变换的土壤光谱反射率与土壤有机质含量相关性显著提高,复垦区土壤光谱经过多元散射校正和一元微分处理并利用偏最小二乘回归模型建模预测效果最好。当前较少有研究对矿区复垦农田土壤有机质进行高光谱估测,本研究成果可为有效利用高光谱遥感技术,快速、有效地测定复垦农田土壤有机质含量提供技术支撑。     

应用可见/近红外高光谱成像测定鲑鱼片脂肪含量分布(英文)

脂肪作为一种重要的品质参数,在大西洋鲑鱼片中的分布很不均匀。为寻找一种能替代脂肪化学检测的快速无损的方法,该研究应用可见/近红外高光谱成像测定大西洋鲑鱼片的脂肪含量分布。分别采用可见/短波近红外(400-1100 nm)和近红外(900-1700 nm)系统获取大西洋鲑鱼片样本的高光谱图像。提取样本图像的平均光谱并与其相应的脂肪含量化学值采用偏最小二乘回归(partial least squares regression,PLSR)和最小二乘支持向量机(least-squares support vector machines,LS-SVM)建立相关性模型。为降低高光谱图像的共线性和冗余度,基于竞争性自适应重加权算法(competitive adaptive reweighted sampling,CARS)分别在可见/短波近红外和近红外光谱区间提取16个(468,479,728,734,785,822,863,890,895,899,920,978,1005,1033,1040,1051 nm)和15个(975,995,1023,1047,1095,1124,1167,1210,1273,1316,1354,1368,1575,1632,1661 nm)特征波长,并分别建立PLSR和LS-SVM模型。特征波长模型的性能优于全波段模型,且近红外区间的特征波长PLSR模型为最优,预测决定系数(R2p)为0.92,预测均方根误差(root mean square error of prediction,RMSEP)为0.92%,剩余预测偏差(residual predictive deviation,RPD)为3.50。最后,将最优模型用于预测高光谱图像上所有像素点的脂肪含量以展示样本上脂肪的分布。此外,还基于该技术对大西洋鲑整鱼片实现了脂肪分布可视化。结果表明高光谱成像技术结合化学计量学方法在大西洋鲑鱼片脂肪的定量和分布可视化上有一定的研究和应用前景。     

基于高光谱图像的茶树LAI与氮含量反演

为了对茶树进行实时、快速、无损的叶面积指数LAI和氮含量检测,该文以英红九号茶树为试验对象,利用便携式高光谱成像仪采集光谱数据、人工破坏性采摘叶片进行叶面积指数的计算以及传统化学方法测量叶片氮含量,比较不同高光谱特征变换形式与LAI和氮含量之间的相关性,并选择其中相关系数较高的高光谱特征变量作为自变量,分别采用线性、指数、对数和抛物线表达式建立LAI和氮含量的回归模型。结果显示:在多种高光谱数据变量建立的模型中,以绿峰反射率R_g为自变量的对数拟合模型最佳,其拟合样本的决定系数R~2和验证样本的均方根误差RMSE值分别为0.9和0.087 6。以植被指数变量VI_4(红边面积/黄边面积)与氮含量建立的指数模型为最佳建模效果,拟合样本的决定系数R~2和验证样本的均方根误差RMSE值分别为0.830 3和0.102 9,研究结果可为茶树叶面积指数LAI和营养成分的无损检测提供参考。     

基于分数阶微分预处理高光谱数据的荒漠土壤有机碳含量估算

对光谱数据进行预处理是提升高光谱建模精度十分必要和有效的途径。为了研究分数阶微分预处理方法在高光谱数据估算荒漠土壤有机碳含量中的应用,该研究以艾比湖流域为研究靶区,利用2015年5月采集的103个表层土壤样本的实测有机碳数据和室内测定的高光谱数据,以0.2阶为步长对原始光谱反射率及对应的倒数变换、对数变换、对数倒数变换、均方根变换的0-2阶微分进行分数阶运算预处理并研究其与土壤有机碳含量相关性,基于通过0.01显著性检验的特征波段对土壤有机碳含量进行偏最小二乘回归建模并进行精度分析。结果表明：1）分数阶微分预处理可以细化土壤有机碳及其光谱反射率相关性的变化趋势;2）各阶微分预处理后的相关系数通过显著性检验波段的数量均呈现先增后减的趋势,但波段数量最多的对应阶数并不统一;3）对数变换的1.6阶微分所建立的模型为最优模型,该模型的RMSEC=2.433 g/kg,R2c=0.786,RMSEP=2.263 g/kg,R2p=0.825,RPD=2.392。说明了分数阶预处理过后的模型精度和稳健性较整数阶微分有了大幅提升,并且运用在高光谱反演土壤有机碳含量上是可行的。     

高光谱成像快速检测壳聚糖涂膜草莓可溶性固形物

为了对壳聚糖涂膜草莓可溶性固形物含量(soluble solids content, SSC)进行快速检测,该文采用高光谱成像仪(400～1 000 nm)对0,0.5%,1%浓度的壳聚糖(chitosan, CTS)涂膜草莓分别储藏1,2,4 d后进行成像,并测量样本SSC。通过分析SSC发现,0.5%和1%壳聚糖涂膜草莓,其SSC随着储藏天数的增加均高于0浓度壳聚糖涂膜草莓,说明了0.5%和1%壳聚糖涂层抑制了草莓中SSC的降低,能够延长草莓的新鲜口味。随后采用蒙特卡罗-偏最小二乘法(monte carlo-partial least squares, MCPLS)对异常样本进行剔除。对剔除异常样本后的光谱数据进行不同预处理,以确定最优的预处理方法。为提高运行速度和降低数据维数,采用竞争性自适应权重取样法(competitive adaptive reweighted sampling, CARS)和连续投影算法(successive projections algorithm, SPA)进行特征波段选择。最后,采用偏最小二乘回归(partial least square regression, PLSR)和支持向量回归(support vector regression, SVR)法建立回归模型。最终结果表明:SPA-SVR模型效果最佳,0浓度的壳聚糖涂膜的草莓,建模集精度Rc2为0.865,预测集精度Rv2为0.835;0.5%浓度的壳聚糖涂膜的草莓,建模集精度Rc2为0.808,预测集精度Rv2为0.799;1%浓度的壳聚糖涂膜的草莓,建模集精度Rc2为0.834,预测集精度Rv2为0.875。对储藏第4天的部分样本图像进行主成分分析(principal component analysis, PCA),结果显示除第二主成分图像(PC2)中有部分噪声影响外,PC1和PC3均能完整反映草莓信息,且PC3图像明显呈现出不同浓度壳聚糖涂膜草莓的褐变程度,说明不同浓度的壳聚糖涂膜也会对草莓货架期产生不同影响。综上说明利用高光谱成像技术可以实现壳聚糖涂膜草莓SSC快速检测,有效指导草莓保鲜处理。     

利用近红外高光谱图像技术快速鉴别西瓜种子品种

为了研究采用近红外高光谱图像技术对西瓜种子品种快速无损鉴别的可行性,该文采用近红外高光谱图像技术,通过提取西瓜种子的光谱反射率,结合Savitzky-Golay (SG)平滑算法,经验模态分解算法（empirical mode decomposition,EMD）和小波分析（wavelet transform,WT）对提取出的光谱数据进行去除噪声处理,采用连续投影算法（successive projections algorithm,SPA）和遗传-偏最小二乘法（genetic algorithm-partial least squares, GA-PLS）进行特征波长选择。基于全波段光谱建立了偏最小二乘判别分析（partial least squares-discriminant analysis,PLS-DA）,基于特征波长建立了反向传播神经网络（back-propagation neural network,BP NN)判别模型和极限学习机（extreme learning machine,ELM）判别模型。试验结果表明,基于特征波长的BPNN模型和ELM模型的结果优于基于全部波长的PLS-DA模型,基于SG预处理光谱提取的特征波长建立的ELM模型取得最优的判别效果,建模集和预测集的判别正确率均为100%。结果表明应用近红外高光谱成像技术对西瓜种子品种鉴别是可行的,为西瓜种子的品种快速鉴别提供了一种新方法。     

Factors affecting the levels of tea polyphenols and caffeine in tea leaves

An isocratic HPLC procedure was developed for the simultaneous determination of caffeine and six catechins in tea samples. When 31 commercial teas extracted by boiling water or 75% ethanol were analyzed by HPLC, the levels of (-)-epigallocatechin 3-gallate (EGCG), and total catechins in teas were in the order green tea (old leaves) > green tea (young leaves) and oolong tea > black tea and pu-erh tea. Tea samples extracted by 75% ethanol could yield higher levels of EGCG and total catechins. The contents of caffeine and catechins also have been measured in fresh tea leaves from the Tea Experiment Station in Wen-Shan or Taitung; the old tea leaves contain less caffeine but more EGCG and total catechins than young ones. To compare caffeine and catechins in the same tea but manufactured by different fermentation processes, the level of caffeine in different manufactured teas was in the order black tea > oolong tea > green tea > fresh tea leaf, but the levels of EGCG and total catechins were in the order green tea > oolong tea > fresh tea leaf > black tea. In addition, six commercial tea extracts were used to test the biological functions including hydroxyl radical scavenging, nitric oxide suppressing, and apoptotic effects. The pu-erh tea extracts protected the plasmid DNA from damage by the Fenton reaction as well as the control at a concentration of 100 microg/mL. The nitric oxide suppressing effect of tea extracts was in the order pu-erh tea >/= black tea > green tea > oolong tea. The induction of apoptosis by tea extract has been demonstrated by DNA fragmentation ladder and flow cytometry. It appeared that the ability of tea extracts to induce HL-60 cells apoptosis was in the order green tea > oolong > black tea > pu-erh tea. All tea extracts extracted by 75% ethanol have stronger biological functions than those extracted by boiling water.     

HPLC analysis of naturally occurring methylated catechins, 3' '- and 4' '-methyl-epigallocatechin gallate, in various fresh tea leaves and commercial teas and their potent inhibitory effects on inducible nitric oxide synthase in macrophages

(-)-Epigallocatechin-3-gallate (EGCG), a major polyphenol of green tea, undergoes substantial biotransformation to species that includes the methylated compounds. Recent studies have demonstrated that the methylated EGCG has many biological activities. In this study, we have investigated the composition of the three O-methylated EGCG derivatives, (-)-epigallocatechin-3-O-(3-O-methyl)gallate (3' '-Me-EGCG), (-)-epigallocatechin-3-O-(4-O-methyl)gallate (4' '-Me-EGCG) and (-)-4'-methyl epigallocatechin-3-O-(4-O-methyl)gallate (4',4' '-di-Me-EGCG) in tea leaves which were picked from various species and at various seasons, ages of leaves, locations, and fermentation levels. Higher levels of 3' '-Me-EGCG and 4' '-Me-EGCG were detected in Chinshin-Kanzai (a species of Camellia sinensis) cultivated in the mountain area of Sansia, Taipei County, Taiwan. Also, these O-methylated EGCG levels were found to be higher in autumn and winter than in spring and summer. The young leaves were found to be richer in the O-methylated compounds than old leaves and the amount of O-methylated EGCG was higher in unfermented longjin green tea than in semifermented oolong tea. However, the fermented black tea and puerh tea did not contain these compounds. 4',4' '-diMe-EGCG could not be detected in either fresh tea leaves or commercial tea leaves. We also found that 3' '-Me-EGCG has a higher inhibitory effect on the nitric oxide generation and inducible nitric oxide synthase (iNOS) expression as compared with EGCG, while 4' '-Me-EGCG and 4',4' '-di-Me-EGCG were less effective.     

不同茶树品种(系)的绿茶滋味分析及评价模型构建

为客观准确地评价不同茶树品种(系)绿茶的滋味品质,该研究系统分析了25个茶树品种(系)绿茶的主要滋味成分含量及其Dot值,利用主成分分析法对不同茶树品种(系)绿茶滋味品质进行综合评价并建立滋味品质评价模型。结果表明:儿茶素、咖啡碱是不同茶树品种(系)绿茶苦涩味强度差异的主要原因,其中表没食子儿茶素没食子酸酯(Epigallocatechingallate,EGCG)是所测样品中涩味的主要贡献物质,EGCG和咖啡碱为所测样品中苦味的主要贡献物质。氨基酸是不同茶树品种(系)绿茶鲜味、甜味差异主要因素,谷氨酸是所测样品中鲜味的主要贡献物质。主成分分析表明:前5个主成分的累计方差贡献率为79.018%,并以前5个主成分的线性回归方程和贡献率构建了滋味品质评价模型,模型评价结果与感官审评结果较为相似,存在极显著相关性(P0.01),相关系数为0.791;利用模型中各主成分的得分,能够较好的区分所测样品的苦、涩、鲜、爽等滋味特征。因此,该研究所建模型能够较好的评价所测样品的滋味品质及其主要滋味特征的差异,为各茶树品种(系)的推广应用提供理论依据,也为不同茶树品种(系)绿茶滋味品质的科学评价提供新的思路和方法。     

便携式茶鲜叶品质光谱检测装置研制

品质监测对茶鲜叶适时采摘和茶叶加工品控具有重要意义。该研究基于可见/近红外光谱技术,研发了便携式茶鲜叶品质无损检测装置。该装置分为主机和手柄2部分,主机大小约240 mm×250 mm×240 mm,包括光谱仪、光源、可充电锂电池、稳压板和散热风扇;手柄大小约130 mm×100 mm×30 mm,包括光纤探头、金属灯杯、白参考板和外触发按钮。基于该设备,采集了茶鲜叶500～900 nm范围内可见/近红外漫反射光谱,对比了归一化（Normalize,NOR）、一阶导数（First Derivative,FD）、标准正态变量变换（Standard Normal Variable Transformation,SNV）和概率商归一化（Probabilistic Quotient Normalization,PQNOR）等不同光谱预处理方法对茶叶光谱的处理结果,建立了茶鲜叶干物质含量、水浸出物含量、茶多酚含量的偏最小二乘定量预测模型。结果表明,PQNOR预处理后建立的偏最小二乘预测模型精度最好,干物质、水浸出物和茶多酚含量预测模型在验证集的相关系数分别为0.905、0.896和0.747,均方根误差分别为0.856%、0.559%和0.549%。在茶园对装置的精度进行了现场测试,单片茶鲜叶检测时间约为1 s,干物质、水浸出物和茶多酚含量预测值与测量值的均方根误差分别为0.903%、0.634%和0.551%。该装置可以实现茶鲜叶光谱原位采集和干物质含量、水浸出物、茶多酚的定量分析,对及时掌握茶树生长情况、辅助决策采茶方案和保障茶叶品质具有重要作用。     

基于IRIV算法优选大豆叶片高光谱特征波长变量估测SPAD值

在植物叶绿素特征波长变量筛选过程中,与叶绿素关系较弱的波长变量极易被忽略,导致这些弱信息变量包含叶绿素的有效信息丢失,因此,确定叶片光谱中弱信息变量对揭示叶绿素高光谱响应规律具有重要意义。该研究以江汉平原大豆鼓粒期的叶片为研究对象,采集80组大豆叶片高光谱和SPAD(soil and plant analyzer development)值,分析SPAD值与大豆叶片反射率相关关系和光谱波长变量自相关关系,基于迭代和保留信息变量法(iteratively retains informative variables,IRIV)筛选大豆叶片的特征波长变量,建立偏最小二乘回归(partial least squares regression,PLSR)和支持向量机(support vector machine,SVM)模型估测SPAD值。结果表明,大豆叶片SPAD值与光谱反射率在可见光波段具有极显著负相关,在近红外波段存在不显著的正相关性(P0.01);可见光、近红外2波段的波长变量之间相关性较弱,但2波段内变量之间的相关性较强;基于IRIV算法确定了大豆叶绿素的特征波长变量,利用特征波长变量建立的估测模型的估测能力高于仅利用强信息波长变量建立的估测模型,表明弱信息变量对估测叶片SPAD值具有重要意义;IRIV-SVM模型估测能力最优,验证集R2和相对分析误差(RPD)分别为0.73、1.82。该文尝试证明了光谱中弱信息变量的重要性,为揭示叶片高光谱响应机理提供了理论依据。     

基于高光谱图像和遗传优化神经网络的茶叶病斑识别

为实现茶叶病害的快速高效识别,提出了基于高光谱成像技术和图像处理技术融合的茶叶病斑识别方法。利用高光谱成像技术采集了炭疽病、赤叶斑病、茶白星病、健康叶片等4类样本的高光谱图像。提取感兴趣区域敏感波段的相对光谱反射率作为光谱特征。通过2次主成分分析,确定第二次主成分分析后的第二主成分图像为特征图像,基于颜色矩和灰度共生矩阵提取特征图像的颜色特征和纹理特征。利用BP神经网络对颜色、纹理和光谱特征向量融合数据进行检验,识别率为89.59%;为提高识别率,提出遗传算法优化BP神经网络的方法,使病斑识别率提高到94.17%,建模时间也缩短至1.7 s。试验结果表明:高光谱成像技术和遗传优化神经网络可以快速准确的实现对茶叶病斑的识别,可为植保无人机超低空遥感病害监测提供参考。     

Inhibitory effects of tea catechins and O-methylated derivatives of (-)-epigallocatechin-3-O-gallate on mouse type IV allergy

The inhibitory effects of tea catechins, the O-methylated derivatives of (-)-epigallocatechin-3-O-gallate (EGCG), and the polyphenol extracts from tea leaves (Camellia sinensis L.) on oxazolone-induced type IV allergy in male ICR mice were investigated. Four major tea catechins and two O-methylated derivatives, (-)-epigallocatechin-3-O-(3-O-methyl)gallate (EGCG3' 'Me) and (-)-epigallocatechin-3-O-(4-O-methyl)gallate (EGCG4' 'Me), showed significant inhibitory effects on mouse type IV allergy after a percutaneous administration at a dose of 0.13 mg/ear. Among tea catechins, the compounds including galloyl moieties, such as EGCG and (-)-epicatechin-3-O-gallate (ECG), showed the strongest inhibitory activities on mouse type IV allergy. The inhibitory activities of EGCG3' 'Me and EGCG4' 'Me were higher than that of EGCG at a dose of 0.05 mg/ear. Polyphenol extract from tea leaves of Benihomare cultivar, which includes EGCG3' 'Me, strongly inhibited mouse type IV allergy after percutaneous administration in comparison with that from Yabukita cultivar, which does not include EGCG3' 'Me, at doses of 0.05 and 0.13 mg/ear. EGCG3' 'Me is thought to contribute, at least in part, to the inhibitory ability of Benihomare tea leaves on mouse type IV allergy. EGCG and the polyphenol extracts from Benihomare and Yabukita tea leaves also inhibited mouse type IV allergy by oral administration at 1 h before the sensitization and at 1 h before the challenge with oxazolone. Therefore, daily intake of tea drinks could have potential to prevent type IV allergy.     

Metabolic dependence of green tea on plucking positions revisited: a metabolomic study

The dependence of global green tea metabolome on plucking positions was investigated through (1)H nuclear magnetic resonance (NMR) analysis coupled with multivariate statistical data set. Pattern recognition methods, such as principal component analysis (PCA) and orthogonal projection on latent structure-discriminant analysis (OPLS-DA), were employed for a finding metabolic discrimination among fresh green tea leaves plucked at different positions from young to old leaves. In addition to clear metabolic discrimination among green tea leaves, elevations in theanine, caffeine, and gallic acid levels but reductions in catechins, such as epicatechin (EC), epigallocatechin (EGC), epicatechin-3-gallate (ECG), and epigallocatechin-3-gallate (EGCG), glucose, and sucrose levels were observed, as the green tea plant grows up. On the other hand, the younger the green tea leaf is, the more theanine, caffeine, and gallic acid but the lesser catechins accumlated in the green tea leaf, revealing a reverse assocation between theanine and catechins levels due to incorporaton of theanine into catechins with growing up green tea plant. Moreover, as compared to the tea leaf, the observation of marked high levels of theanine and low levels of catechins in green tea stems exhibited a distinct tea plant metabolism between the tea leaf and the stem. This metabolomic approach highlights taking insight to global metabolic dependence of green tea leaf on plucking position, thereby providing distinct information on green tea production with specific tea quality.