光谱指数的植物叶片叶绿素含量估算模型

引用本文

李哲, 张飞, 陈丽华, 张海威. 光谱指数的植物叶片叶绿素含量估算模型[J]. 光谱学与光谱分析, 2018,38(5): 1533-1539.
LI Zhe, ZHANG Fei, CHEN Li-hua, ZHANG Hai-wei. Research on Spectrum Variance of Vegetation Leaves and Estimation Model for Leaf Chlorophyll Content Based on the Spectral Index[J]. Spectroscopy and Spectral Analysis, 2018,38(5): 1533-1539.
Doi:10.3964/j.issn.1000-0593(2018)05-1533-07 复制到剪切板

Permissions

《光谱学与光谱分析》期刊社所有

光谱指数的植物叶片叶绿素含量估算模型

李哲^1,², 张飞^1,^2,^3,^*, 陈丽华⁴, 张海威^1,²

1. 新疆大学资源与环境科学学院, 新疆乌鲁木齐 830046

2. 新疆大学绿洲生态教育部重点实验室, 新疆乌鲁木齐 830046

3. 新疆智慧城市与环境建模普通高校重点实验室, 新疆乌鲁木齐 830046

4. 新疆艾比湖湿地国家级自然保护区管理局, 新疆博乐 833400

*通讯联系人 e-mail: zhangfei3s@163.com

作者简介: 李哲, 女, 1993年生, 新疆大学资源与环境科学学院硕士研究生 e-mail: 13999940867@163.com

收稿日期: 2017-05-02

基金: 国家自然科学基金项目-新疆本地优秀青年培养专项(U1503302), 国家自然科学基金项目(41361045)资助

摘要

叶片叶绿素能够有效监测植被的生长状况, 利用光谱指数反演植被叶绿素含量是目前的通用方法。实测了盐生植物光谱反射率和叶片叶绿素含量。对SPAD值进行变换, 对比Pearson与VIP方法探讨盐生植被叶片叶绿素含量与植被指数的相关性并进行精度验证, 从中选出最佳拟合模型。研究表明, 通过对Pearson与VIP相关性分析, 最终选定VIP方法建立植被指数的叶片叶绿素估算模型, NDVI₇₀₅, ARVI, CI_{red edge}, PRI, VARI, PSRI和NPCI的VIP值均大于0.8, 因此选定这七个植被指数为最优植被指数; 预测结果显示, 所有模型的相关性都在0.7以上, 预测值与实测值相关性最好的是经过倒数变换的SPAD值, R=0.816, RMSE=0.007。基于VIP方法的反演模型能较好地估算研究区植被叶绿素含量, 该方法为植物叶绿素含量诊断的实际应用提供了重要的理论依据和技术支持。

关键词: 叶绿素含量; 植被指数; 估算模型

中图分类号:TP79 文献标志码:A

Research on Spectrum Variance of Vegetation Leaves and Estimation Model for Leaf Chlorophyll Content Based on the Spectral Index

LI Zhe^1,², ZHANG Fei^1,^2,^3,^*, CHEN Li-hua⁴, ZHANG Hai-wei^1,²

1. College of Resources & Environmental Science, Xinjiang University, Urumqi 830046, China;

2. Key Laboratory of Oasis Ecology, Xinjiang University, Urumqi 830046, China

3. General Institutes of Higher Learning Key Laboratory of Smart City and Environmental Modeling, Xinjiang University, Urumqi 830046, China

4. Area Management Bureau of Ebinur Lake Wetland Natural Reserve, Bole 833400, China

Abstract

The chlorophyll can effectively monitoring vegetation growth status, currently the hyperspectral vegetation index (VI) is one of the common methods that have been widely applied to estimate the leaf chlorophyll content (LCC) inversion. Non-destructive rapid estimation of chlorophyll using hyperspectral remote sensing technology is a prerequisite to dynamically monitor chlorophyll content, and it is an important research issue of vegetation remote sensing. The author measured the leaf spectral reflectance and chlorophyll relative content of desert plants, analyzed the spectral curves of different desert plants under the same chlorophyll content, then transformed the SPAD value, compared Pearson and VIP methods to study the correlation between chlorophyll content and vegetation index of desert vegetation. Finally, the author selected the best fitting model from accuracy test. The results showed that: Based on the comparative analysis between Pearson and VIP, established the chlorophyll content estimation model by VIP method, selected 7 vegetation indices, which was NDVI₇₀₅, ARVI, CI_{red edge}, PRI, VARI, PSRI and NPCI respectively, the value of VIP all greater than 0.8, thus these 7 vegetation indices were the optimal vegetation indices. The prediction results indicated that the correlation of all models was more than 0.7, the best correlation between the predicted value and the measured value was the SPAD value of the reciprocal transformation, R=0.816, RMSE=0.007. The inversion model based on VIP method can estimate the chlorophyll content of vegetation in the study area, it provides an important theoretical basis and technical support for the practical application in the diagnosis of plant chlorophyll content.

Keyword: Chlorophyll content; Vegetation index; Estimation model

文章图片

引言

叶绿素是植物生长过程的一个重要生化参数, 是植物光合作用能力的指示器, 已被广泛用于植物的养分循环、发育情况、生产力和疾病监测等研究^[1], 叶片叶绿素含量(leaf chlorophyll content, LCC)能够反映植物的生长状况^[2], 因此, 在生态系统较为脆弱的西北荒漠地带, 监测植被叶片的叶绿素含量对干旱区生态系统的恢复与重建和生态系统生产力的评估具有重要指导意义。

Le Maire等^[3]总结了2002年之前所有发表的有关叶绿素含量预测的光谱指数。在电磁波的蓝光和红光范围, 叶绿素表现为强烈吸收, 光谱指数基于那些对叶绿素敏感而对其他因素不敏感的窄波段光谱比值构建。应用于叶片尺度的植被光谱指数大多数结构相对简单, 常用比值型、差值型、归一化差值型、新二重差值型及改进版本等。较为典型的有改进的叶绿素吸收指数MCARI和土壤调节指数, 其优点在于消除冠层参数, 如叶面积指数LAI和土壤背景变化等影响^[4]。 Dash等^[5]针对高浓度的植被叶绿素含量, 发展了新型植被指数MTCI; Daughtry等^[6]根据前人经验, 并综合OSAVI指数的优点, 构建MCARI指数的同时, 进一步提出了MCARI/OSAVI和TCARI/OSAVI指数, 提高了反演叶绿素的准确性; 李鑫川等利用可见光-近红外与短波红外的特征建立了新型植被指数来估算农作物的叶面积指数; 廖钦洪等^[7]有效利用遥感影像波段对各观测角的玉米冠层反射率进行分析, 对比新型植被指数HD-TCARI和TCARI反演叶绿素的特点。

由于受西北干旱区水盐胁迫的影响, 使得研究区的典型植被如柽柳(Tamarix)、梭梭(Haloxylon ammodendron)、白刺(Nitraria tangutorum)等与普通的植被相比形态和生理特征都具有差异性。本工作选取多种已发表的叶绿素植被指数, 通过数学变换手段研究这些指数在盐生植被特定季节的适应性, 并对结果进行分析评价, 从中筛选出适应荒漠植被特定季节的几种光谱指数。

1 实验部分

1.1 采样区域

艾比湖湿地自然保护区位于新疆博尔塔拉蒙古自治州(44° 30'— 45° 09'N, 82° 36'— 83° 50'E), 与精河县、阿拉山口市和博乐市相邻。总面积2 670.85 km², 其中水域面积为683.81 km²。艾比湖湿地是我国典型的沙漠湖泊湿地, 湖边植物具有良好的防风固沙的特点, 保护区夏季干旱少雨, 冬季寒冷, 年日照时数大于2 500 h, 年均温度8.3 ℃, 年均降水90 mm, 蒸发量高于3 000 mm, 属于中温带大陆性干旱气候^[8]。保护区内有各类盐生植物占全国盐生植物种类的60%左右, 物种多样性低, 优势种主要为耐盐碱性的梭梭、柽柳、盐穗木(Halostachys caspica)、盐节木(Halocnemum strobilaceum)、花花柴(Karelinia caspica)、芦苇(Phragmites australis)等。

1.2 植物光谱测定

于2016年10月对艾比湖湿地国家自然保护区进行实地考察, 保护区布设35个样点, 其中每个样点采集两种荒漠植被。用新鲜的冰块对采集回来的样本进行冷却保存, 并放入密封的塑料袋, 确保植被样本不被损坏。使用便携式地物光谱(Field Spec HiRes, ASD, 美国)对植物光谱曲线进行采集, 该仪器光谱范围为350~2 500 nm, 光谱分辨率3 nm。测量前用白板定标, 仪器距样点上方5 cm, 每个点位进行5次反射光谱测量, 最后取其平均值作为该点的原始光谱反射率。然后利用光谱仪的后处理软件分析采集的植物光谱曲线, 将每个点号的5条光谱数据置于Process下Statistics内的Mean命令进行计算平均值, 然后在Origin8.0软件中进行微分变换与多元统计分析。

1.3 叶片叶绿素测定

采用日本KONICA MINOLTA SPAD-502叶绿素计对盐生植物叶片(梭梭、胡杨(Populus euphratica)、柽柳、芦苇、白刺、花花柴、盐节木)的相对叶绿素含量快速无损监测。通过测量叶片的吸收率计算出SPAD值, 得到的数据应与叶片内部结构的叶绿素浓度成正比^[9]。通常情况下, 叶片叶绿素浓度的相对值就是SPAD的值。盐生植物叶片的SPAD值与光谱测量同步进行, 每个样点选取2种植被, 每种四棵植株, 每株植被选择健康未枯萎, 并且是第2、 3片叶片进行测量; 每叶片测得10个SPAD值, 最后取其平均值作为该叶片的SPAD值。

2 研究方法

2.1 高光谱植被指数的选择

研究^[10]表明, 光谱指数和叶绿素含量之间的经验模型是田间和实验室测量叶绿素含量进行验证的有效估算方法。因此, 选取了15个前人构建的高光谱植被指数(表1)应用于荒漠植被叶片叶绿素含量的反演研究。

表1 光谱指数计算公式 Table 1 Formula for the spectral index

2.2 Pearson相关

Pearson相关系数, 通常用来度量两个变量X和Y之间的相互关系, 计算公式为^[17]

$ρ = Cor (X, Y) = \frac{Cov (X, Y)}{\sqrt[]{Var (X) Var (Y)}} (1)$

式中, Cov(X, Y)代表X与Y的协方差, Var(X)和Var(Y)代表X和Y的方差。当相关系数为1时, X与Y的关系为Y=aX+b, a> 0; 当相关性为-1时, X与Y的关系为Y=aX+b, a< 0。如果X与Y相互独立, 相关系数为0。

2.3 变量投影重要性(VIP)

通过计算变量投影重要性指标VIP来分析各高光谱植被指数对盐生植被叶片叶绿素浓度预的贡献。 VIP的计算公式为

$V_{VI P_{j}} = \sqrt[]{\frac{p \overset{m}{\sum_{h = 1}} R_{d} (Y, t_{h}) w_{h j}^{2}}{\overset{m}{\sum_{h = 1}} R_{d} (Y, t_{h})}} (2)$

式中, $V_{VI P_{j}}$ 表示第j个自变量的投影相关程度; p为自变量数量; m为基于PLSR的成分提取数量; t_h代表第h个成分(h=1, 2, 3, …, m); R_d(Y, t_h)表示成分t_h对因变量Y的解释程度; W_hj是轴W_h的第j个分量, w_h为矩阵 $X_{h - 1}^{T}$ Y_h_-1 ${Y^{T}}_{h - 1}$ X_h_-1的特征向量。如果自变量x_j解释因变量Y的作用均相同, 其对应的 $V_{VI P_{j}}$ 值均为1, 对应较高 $V_{VI P_{j}}$ 值的自变量x_j, 其在解释因变量Y的作用越重要^[18]。

3 结果与讨论

3.1 Pearson与VIP相关性分析对比

将SPAD值(R)进行了LOG(R), SQRT(R), 1/R, LOG(1/R), SQRT(1/R), 1/LOG(R)六种数学变换, 对比两种相关分析方法以研究其相关性, 然后选用相关性最好的一种方法建立叶片叶绿素含量估算模型(n=70), 并随机选取23个独立样本进行预测验证。首先采用Pearson相关分析分析研究多种变换下SPAD值与植被指数的相关性(见图1)。

由图1可知, 在未变换的Pearson相关矩阵图中, 与SPAD值呈现较好相关性的共有4个植被指数, 相关性由大到小分别是: PSRI(-0.597^{* *}), NPCI(-0.535^{* *}), VARI(0.494^*)和ARVI(0.421^*), 其中PSRI和NPCI在0.01水平上显著相关, VARI和ARVI在0.05水平上显著相关。在其余变换的Pearson相关矩阵图中, 与SPAD值呈现较好相关性的共有5个植被指数, 分别是PSRI, NPCI, VARI, ARVI和PRI。在LOG(R)变换的Pearson相关矩阵图中, 相关性由大到小分别是: PSRI(-0.634^{* *}), NPCI(-0.577^{* *}), VARI(0.562^{* *}), ARVI(0.441^*)和PRI(-0.433^*), 其中PSRI, NPCI, VARI在0.01水平上显著相关, ARVI和PRI在0.05水平上显著相关。在SQRT(R)变换的Pearson相关矩阵图中, 相关性由大到小分别是: PSRI(-0.620^{* *}), NPCI(-0.560^{* *}), VARI(0.531^{* *}), ARVI(0.435^*)和PRI(-0.426^*), 其中PSRI, NPCI, VARI在0.01水平上显著相关, ARVI和PRI在0.05水平上显著相关。在1/R变换的Pearson相关矩阵中, 相关性由大到小分别是: PSRI(0.629^{* *}), VARI(-0.595^{* *}), NPCI(0.585^{* *}), ARVI(-0.425^*)和PRI(0.418^*), 其中PSRI, VARI和NPCI在0.01水平上显著相关, ARVI和PRI在0.05水平上显著相关。在LOG(1/R)变换Pearson相关矩阵图中, 相关性由大到小分别是: PSRI(0.634^{* *}), NPCI(0.577^{* *}), VARI(-0.562^{* *}), ARVI(-0.441^*)和PRI(0.433^*), 其中PSRI, VARI和NPCI在0.01水墙上显著相关, ARVI和PRI在0.05水平上显著相关。在SQRT(1/R)变换Pearson相关矩阵图中, 相关性由大到小分别是: PSRI(0.637^{* *}), VARI(0.585^{* *}), NPCI(-0.583^{* *}), ARVI(-0.437^*)和PRI(0.430^*), 其中PSRI, VARI和NPCI在0.01水平上显著相关, ARVI和PRI在0.05水平上显著相关。在1/LOG(R变换Pearson相关矩阵图中, 相关性由大到小分别是: PSRI(0.635^{* *}), VARI(-0.585^{* *}), NPCI(0.585^{* *}), ARVI(-0.434^*)和PRI(0.427^*), 其中PSRI, VARI和NPCI在0.01水平上显著相关, ARVI和PRI在0.05水平上显著相关。

	Figure Option View Download New Window
	图1 Pearson相关矩阵图Fig.1 Correlation matrix of Pearson

通过Pearson相关性分析, 筛选出了五个植被指数, 分别是PSRI, NPCI, VARI, ARVI和PRI。显著性在0.01水平的植被指数分别是PSRI, VARI和NPCI, 表明该三种植被指数与SPAD值的相关性较好。将变换前后的SPAD值与筛选的植被指数用SPSS软件进行多元回归线性分析, 得到如下结果(见表2)。

表2 SPAD值与植被指数拟合评定指标(Pearson) Table 2 The fit evaluation indicators of the SPAD with vegetation index (Pearson)

笔者在Pearson相关分析评定指标结果的基础上, 又采用了VIP相关分析研究多种变换下SPAD值与植被指数的相关性(见图2)。

	Figure Option View Download New Window
	图2 VIP图Fig.2 The picture of VIP

基于光谱测量数据和PLSR模型的VIP方法计算分析结果, 显示不同SPAD变换应用中, 各植被光谱指数在估算盐生植物叶片叶绿素时的重要性(图2)。利用VIP方法进行重要性评价时, 当VIP值大于1时, 说明该植被指数在估算植物叶片叶绿素时具有重要作用; 当植被指数的VIP值在0.5~1之间时, 说明该植被指数的预测重要性还需要进一步分析, 需要通过增加样本或根据其他条件进行判断分析, 而一般情况下, VIP值大于0.8的对应植被指数是预测叶片叶绿素值的重要植被指数^[19]。

从图2中可以看出, NDVI₇₀₅, ARVI, CI_{red edge}, PRI, VARI, PSRI和NPCI这七个指数的VIP值始终大于0.8, 其中ARVI, CI_{red edge}, PRI, VARI, PSRI和NPCI的值始终大于1, 说明这些植被指数在估算植物叶片叶绿素时具有重要作用。因此, 根据VIP图选取了不同的植被指数, 将变换前后的SPAD值与筛选的植被指数用SPSS软件进行多元回归线性分析, 得到如表3所示结果。

表3 SPAD值与植被指数拟合评定指标(VIP) Table 3 The fit evaluation indicators of the SPAD with vegetation index (VIP)

由表3得知, 1/R, SQRT(1/R)和1/(LOGR)这三种SPAD变换与对应筛选的植被指数相关性最好, 相关性达到0.8, 根据R最大, RMSE最小的原则, 相关性最优的叶绿素变换是SPAD倒数变换, 相关性最小的是未经变换的SPAD值, R仅为0.73, RMSE最大达到11.28。

3.2 基于VIP的盐生植物叶片叶绿素估算模型

通过对比Pearson与VIP的SPAD值与植被指数的拟合评定指标表, 最终采用VIP筛选的植被指数建立叶片叶绿素的估算模型(见表4)。

表4 SPAD值与植被指数最佳估算模型 Table 4 The optimal estimation model of SPAD and vegetation index

3.3 模型的预测与验证

为了进一步验证模型的准确性和稳定性, 采用预测集的23个样本对叶绿素含量估算模型的预测能力进行验证(图3)。

	Figure Option View Download New Window
	图3 预测值与实测值的散点图Fig.3 Scatter plot between the observed LCC and the predicted values with different vegetation indices

从图3中可知, 经过倒数变换的SPAD值其预测值与实测值相关性最好, R达到0.816, RMSE为0.007, 所有模型的相关性都在0.7以上, 模型相对较差的是未经变换的SPAD值, R为0.742, RMSE为6.36。可以看出, 基于不同植被指数的反演模型都能较好地得到研究区植被叶绿素含量, 经过验证发现这七种模型对盐生植被叶绿素含量的预测是可行的。

4 结论

采用野外实测光谱及其室内实测叶绿素数据, 对SPAD值进行数学变换, 探索高光谱遥感识别植被叶片叶绿素含量的敏感波段, 通过对比验证Pearson与VIP模型的优劣, 探索盐生植被叶绿素含量与植被指数的相关性, 建立最佳估算模型, 并对模型进行了检验和预测, 这是一个新的尝试。本研究仍有些许不足之处, 受时间和自然条件限制, 采集的样本数量较少, 需要适当增加样本数量。结果利用利用高光谱植被指数反演秋季的盐生植被叶绿素是可行的, 但对于全年的植被叶绿素含量反演仍需要进一步的研究探讨, 以提高光谱指数的普适性。

通过对比作者最终采用VIP筛选的植被指数建立叶片叶绿素的估算模型, 筛选了NDVI₇₀₅, ARVI, CI_{red edge}, PRI, VARI, PSRI和NPCI七个指数。基于VIP方法的SPAD值与植被指数的预测模型相关性都在0.7以上, 根据R最大RMSE最小的原则, 预测值与实测值相关性最好的是经过倒数变换的SPAD值, R达到0.816, RMSE为0.007。因此基于VIP方法的反演模型被认为是最佳估算荒漠植被叶片叶绿素含量模型。

The authors have declared that no competing interests exist.

参考文献

文献列表

[1]	Yang X G, Yu Y, Fan W Y. Environmental Monitoring Assessment, 2015, 187(7): 1. [本文引用:1]
[2]	FENG Wei, ZHU Yan, TIAN Yong-chao, et al(冯伟, 朱艳, 田永超, 等). Acta Ecologica Sinica(生态学报), 2008, 28(10): 4902. [本文引用:1]
[3]	Le Maire G, Francois C, Dufrene E. Remote Sensing of Environment, 2004, 89(1): 1. [本文引用:1]
[4]	ZHANG Si-nan, WANG Quan, JIN Jia, et al(张思楠, 王权, 靳佳, 等). Arid Zone Research(干旱区研究), 2016, 33(5): 1088. [本文引用:1]
[5]	Dash J, Curran P J. IEEE International Symposium on Geoscience and Remote Sensing(IGARSS), 2004, 39(1): 254. [本文引用:1]
[6]	Daughtry C, Walthall C L, Kim M S, et al. Remote Sensing of Environment, 2000, 74(2): 229. [本文引用:1]
[7]	LIAO Qin-hong, ZHANG Dong-yan, WANG Ji-hua, et al(廖钦洪, 张东彦, 王纪华, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2014, 34(6): 1599. [本文引用:1]
[8]	YANG Ai-xia, DING Jian-li, LI Yan-hong, et al(杨爱霞, 丁建丽, 李艳红, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2016, 36(3): 691. [本文引用:1]
[9]	Uddling J, Gelang-Alfredsson J, Piikki K, et al. Photosynthesis Research, 2007, 91(1): 37. [本文引用:1]
[10]	Maccioni A, Agati G, Mazzinghi P. Journal of Photochemistry and Photobiology, 2001, 61(1-2): 52. [本文引用:1]
[11]	Rouse Jr J W, Haas R H, Schell J A, et al. NASA Special Publication, 1974, 351: 309. [本文引用:1]
[12]	Dian Y Y, Le Y, Fang S H, et al. J. Indian Society of Remote Sensing, 2016, 44(4): 583. [本文引用:1]
[13]	TIAN Jing-guo, WANG Shu-dong, ZHANG Li-fu, et al(田静国, 王树东, 张立福, 等). Science Technology and Engineering(科学技术与工程), 2016, 16(15): 1671. [本文引用:1]
[14]	GUO Ni(郭铌). Arid Meteorology(干旱气象), 2003, 21(4): 71. [本文引用:1]
[15]	Francisco F S, John M K, Francisco F V. Wetland s Ecol. Manage, 2013, 21(3): 193. [本文引用:1]
[16]	JIA Kun, YAO Yun-jun, WEI Xiang-qin, et al(贾坤, 姚云军, 魏香琴, 等). Advances in Earth Science(地球科学进展), 2013, 28(7): 774. [本文引用:1]
[17]	ZHANG Jian-yong, GAO Ran, HU Jun, et al(张建勇, 高冉, 胡骏, 等). Journal of Chifeng University·Natural Science Edition(赤峰学院学报·自然科学版), 2014, 30(11): 1. [本文引用:1]
[18]	Svante W, Michael S, Lennart E. Chemometrics and Intelligent Laboratory Systems, 2001, 58(2): 109. [本文引用:1]
[19]	WANG Li-wen, WEI Ya-xing(王莉雯, 卫亚星). Scientia Geographica Sinica(地理科学), 2016, 36(1): 135. [本文引用:1]

2015

0.0

... 引言叶绿素是植物生长过程的一个重要生化参数, 是植物光合作用能力的指示器, 已被广泛用于植物的养分循环、发育情况、生产力和疾病监测等研究^[1], 叶片叶绿素含量(leaf chlorophyll content, LCC)能够反映植物的生长状况^[2], 因此, 在生态系统较为脆弱的西北荒漠地带, 监测植被叶片的叶绿素含量对干旱区生态系统的恢复与重建和生态系统生产力的评估具有重要指导意义 ...

2008

0.0

FENG

Wei

, ZHU

Yan

, TIAN

Yong-chao

, et al(冯伟, 朱艳, 田永超, 等). Acta Ecologica Sinica(生态学报), 2008, 28(10): 4902.

Canopy leaf pigment status is a key index for evaluating crop potential photosynthetic efficiency and nutritional stress. Leaf pigment density per unit ground area provides a rapid and non-destructive method to evaluate yield predictions and management of nitrogen applications. This study investigated the quantitative relationships of leaf pigment density to canopy hyperspectral reflectance in wheat (Tritium aestivum L.) with two field experiments consisting of different cultivars and nitrogen levels in two growing seasons. On the basis of measured protein content, the cultivars, Yumai 34, Yangmai 12 and Ningmai 9 were considered as high, medium and low protein types, respectively. Four nitrogen rates were applied as 0, 75, 150, 225 kg N hm-2 in the form of urea in 2004-2005, and 0, 90, 180, 270 kg N hm-2 in the form of urea in 2005-2006. Canopy hyperspectral reflectance over 350-2500 nm was measured, and the density of chlorophyll a (Chl a), chlorophyll b (Chl b), chlorophyll a+b (Chl a+b), and carotenoid (Car) were determined for leaves in the wheat canopies. The densities of the different pigments (Chl a, Chl b, Chl a+b, Car) in wheat leaves increased with increasing nitrogen application rates. With respect to plant development, the densities of the different pigments (Chl a, Chl b, Chl a+b, Car) initially increased and then decreased after jointing with differences between the cultivars, while gradually decreasing in the non-nitrogen treatment. The sensitive wavebands for pigment density occurred mostly within the visible light range, and a close correlation existed between the first derivatives of reflectance in the red-edge region (680-760 nm) and leaf pigment density. The analyses between 21 vegetation indices and leaf pigment density indicated that the pigment densities were highly correlated to all these 21 vegetation indices. The correlations of leaf chlorophyll density to the following five vegetation indices, were higher than for the other spectral parameters, with the coefficients of determination (R2) for linear correlations above 0.858. These five vegetation indices were Vogelmann indices 2 (Vog2), Vogelmann indices 3 (Vog3), the ratio vegetation index of 810 to 560 nm (RVI(810,560)), the ratio of the red-edge (680-760 nm) slope to the blue-edge (490-530 nm) slope (SRE/SBE), and the ratio of the integral of the first derivative on the red-edge region to that on the blue-edge region (SDr/SDb). All the values of R2 between Car density and the different spectral indices were below 0.780. Testing of the derived regression equations with independent datasets indicated that Vog2, Vog3, SRE/SBE and SDr/SDb were the best to predict leaf pigment density with a relative error (RE, (simulated value - observed value)/observed value) below 17.6%. The R2 associated with the Chl b density was below 0.804, lower than that of other pigment densities. The overall results suggested that the Vog2, Vog3, SRE/SBE and SDr/SDb indices have stable relationships with the pigment density in wheat leaves, especially the densities of leaf Chl a and Chl a+b.

作物叶片色素状况是评价植株光合效率和营养胁迫的重要指标，冠层叶片色素密度(单位土地面积叶片色素总量)的实时无损监测对作物生长诊断、产量估算及氮素管理具有重要意义。以包括不同品质类型（高蛋白、中蛋白和低蛋白）的多个小麦品种在不同施氮水平下的连续2a大田试验为基础，研究了小麦叶片色素密度与冠层高光谱参数的定量关系。结果表明，叶片色素（叶绿素a、叶绿素b、叶绿素a+b和类胡萝卜素）密度随施氮水平增加而提高，不施氮处理的叶片色素密度随生育进程而下降，施氮处理的叶片色素密度呈单峰曲线，品种间存在明显差异。群体叶片色素密度的敏感波段主要分布在可见光区，而红边区域导数光谱表现更显著。光谱参数VOG2、VOG3、RVI(810,560)、SRE/SBE和SDr/SDb等与叶绿素密度关系较为密切，线性方程决定系数R2均在0.858以上，而与类胡萝卜素密度关系减弱，决定系数R2低于0.780，且参数间差异较小。经独立试验资料的检验表明，VOG2、VOG3、SRE/SBE和SDr/SDb对不同色素的估测结果较好，预测相对误差RE低于17.6%，虽然对叶绿素b的准确性稍低。总体上，光谱参数VOG2、VOG3、SRE/SBE和SDr/SDb与小麦群体叶片色素密度关系密切，特别是对叶片叶绿素a和叶绿素a+b的密度可以进行准确可靠的实时监测。

2004

0.0

... Le Maire等^[3]总结了2002年之前所有发表的有关叶绿素含量预测的光谱指数 ...

2016

0.0

... 较为典型的有改进的叶绿素吸收指数MCARI和土壤调节指数, 其优点在于消除冠层参数, 如叶面积指数LAI和土壤背景变化等影响^[4] ...

2004

0.0

... Dash等^[5]针对高浓度的植被叶绿素含量, 发展了新型植被指数MTCI ...

2000

0.0

... Daughtry等^[6]根据前人经验, 并综合OSAVI指数的优点, 构建MCARI指数的同时, 进一步提出了MCARI/OSAVI和TCARI/OSAVI指数, 提高了反演叶绿素的准确性 ...

2014

0.0

LIAO

Qin-hong

, ZHANG

Dong-yan

, WANG

Ji-hua

, et al(廖钦洪, 张东彦, 王纪华, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2014, 34(6): 1599.

The fast estimation of chlorophyll content is significant for understanding the crops growth, monitoring the disease and insect, and assessing the yield of crops. This study gets the hyperspectral imagery data by using a self-developed multi-angular acquisition system during the different maize growth period, the reflectance of maize canopy was extracted accurately from the hyperspectral images under different view angles in the principal plane. The hot-dark-spot index (HDS) of red waveband was calculated through the analysis of simulated values by ACRM model and measured values, then this index was used to modify the vegetation index (TCARI), thus a new vegetation index (HD-TCARI) based on the multi-angular observation was proposed. Finally, the multi-angular hyperspectral imagery data was used to validate the vegetation indexes. The result showed that HD-TCARI could effectively reduce the LAI effects on the assessment of chlorophyll content. When the chlorophyll content was greater than 30 mu g.cm(-2), the correlation (R-2) between HD-TCARI and LAI was only 26.88%similar to 28.72%. In addition, the HD-TCARI could resist the saturation of vegetation index during the assessment of high chlorophyll content. When the LAI varied from 1 to 6, the linear relation between HD-TCARI and chlorophyll content could be improved by 9% compared with TCARI. The ground validation of HD-TCARI by multi-angular hyperspectral image showed that the linear relation between HD-TCARI and chlorophyll content (R-2=66.74%) was better than the TCARI (R-2=39.92%), which indicated that HD-TCARI has good potentials for estimating the chlorophyll content.

[Liao Qin-hong; Zhang Dong-yan] Anhui Univ, Key Lab Intelligent Comp & Signal Proc, Minist Educ, Hefei 230039, Peoples R China.;[Liao Qin-hong; Zhang Dong-yan; Wang Ji-hua; Yang Gui-jun; Yang Hao] Beijing Acad Agr & Forestry Sci, Beijing Agr Informat Technol Res Ctr, Beijing 100097, Peoples R China.;[Liao Qin-hong; Zhang Dong-yan; Wang Da-cheng] Chinese Acad Sci, Inst Remote Sensing & Digital Earth, Beijing 100094, Peoples R China.;[Craig, Coburn; Wong Zhijie] Univ Lethbridge, Dept Geog, Lethbridge, AB T1K 3M4, Canada.

... 廖钦洪等^[7]有效利用遥感影像波段对各观测角的玉米冠层反射率进行分析, 对比新型植被指数HD-TCARI和TCARI反演叶绿素的特点 ...

2016

0.0

... 3 ℃, 年均降水90 mm, 蒸发量高于3 000 mm, 属于中温带大陆性干旱气候^[8] ...

2007

0.0

... 通过测量叶片的吸收率计算出SPAD值, 得到的数据应与叶片内部结构的叶绿素浓度成正比^[9] ...

2001

0.0

... 1 高光谱植被指数的选择研究^[10]表明, 光谱指数和叶绿素含量之间的经验模型是田间和实验室测量叶绿素含量进行验证的有效估算方法 ...

1974

0.0

2016

0.0

2016

0.0

2003

0.0

2013

0.0

2013

0.0

JIA

Kun

, YAO

Yun-jun

, WEI

Xiang-qin

, et al(贾坤, 姚云军, 魏香琴, 等). Advances in Earth Science(地球科学进展), 2013, 28(7): 774.

Fractional vegetation cover is a key parameter for characterizing land surface vegetation coverage, and plays an important role in global change research, earth surface processes simulation and hydro-ecological models. Remote sensing can provide the vegetation coverage information and variation trend on different spatial scales, and become the important means in obtaining the information of regional or global fractional vegetation cover. In this paper, the commonly used remote sensing data sources including hyperspectral data, multispectral data, microwave data and LiDAR data for fractional vegetation cover estimation are analyzed, and multispectral data will be the long term main data source for fractional vegetation cover estimation because of the advantage of its easily acquisition, wide coverage and continuous observation advantages. Then, characteristics and advantages of different estiation methods are analyzed, which include the regression model, the pixel unmixing model, the machine learning method, the physical model, the spectral gradient difference and the forest canopy density mapping model. The dimidiate pixel model of the pixel unmixing model is extensively used for its simple form and certain physical significance, and the neural network method is widely used for generating products because of its extendibility and rapidity of calculation. Furthermore, the existing remote sensing data based fractional vegetation cover products are presented. However, the validation results indicate that almost all of the fractional vegetation cover products have underestimation problem and variational estimation accuracy in different regions. Finally, the fractional vegetation cover estimation research prospect is discussed and high spatial-temporal resolution global fractional vegetation cover dataset, multi-source remote sensing data fusion and assimilation method are the future development of fractional vegetation cover estimation using remote sensing data.

1. College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China; 2. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China

植被覆盖度是刻画地表植被覆盖的重要参数，在全球变化研究、地表过程模拟和水文生态模型中发挥着重要作用。遥感能够反映不同空间尺度的植被覆盖信息及其变化趋势，是获取区域及全球植被覆盖度参数的一个重要手段。综合分析了用于植被覆盖度估算的遥感数据源，包括高光谱数据、多光谱数据、微波数据和激光雷达数据。而且分析了各种常用的植被覆盖度遥感估算方法及其优缺点，并评价了现有基于遥感数据的植被覆盖度产品及存在问题。最后，针对目前研究中存在的问题，讨论了植被覆盖度遥感估算研究的发展趋势，指出高时空分辨率长时间序列的全球植被覆盖度数据集、多源遥感数据融合和同化技术是未来植被覆盖度遥感估算研究的主要方向。 

2014

0.0

... 2 Pearson相关Pearson相关系数, 通常用来度量两个变量X和Y之间的相互关系, 计算公式为^[17] ...

2001

0.0

... 如果自变量x_j解释因变量Y的作用均相同, 其对应的 VVIPj值均为1, 对应较高 VVIPj值的自变量x_j, 其在解释因变量Y的作用越重要^[18] ...

2016

0.0

... 8的对应植被指数是预测叶片叶绿素值的重要植被指数^[19] ...