基于高光谱分析的玉米叶片氮含量分层诊断研究

引用本文

张银杰, 王磊, 白由路, 杨俐苹, 卢艳丽, 张静静, 李格. 基于高光谱分析的玉米叶片氮含量分层诊断研究[J]. 光谱学与光谱分析, 2019,39(9): 2829-2835.
ZHANG Yin-jie, WANG Lei, BAI You-lu, YANG Li-ping, LU Yan-li, ZHANG Jing-jing, LI Ge. Nitrogen Nutrition Diagnostic Based on Hyperspectral Analysis about Different Layers Leaves in Maize[J]. Spectroscopy and Spectral Analysis, 2019,39(9): 2829-2835.
Doi:10.3964/j.issn.1000-0593(2019)09-2829-07 复制到剪切板

Permissions

《光谱学与光谱分析》期刊社所有

基于高光谱分析的玉米叶片氮含量分层诊断研究

张银杰, 王磊^*, 白由路, 杨俐苹, 卢艳丽, 张静静, 李格

农业农村部植物营养与肥料重点实验室, 中国农业科学院农业资源与农业区划研究所, 北京 100081

*通讯联系人 e-mail: wanglei02@caas.cn

作者简介: 张银杰, 女, 1993年生, 中国农业科学院农业资源与农业区划研究所硕士研究生 e-mail: zzhangyinjie@126.com

收稿日期: 2018-07-26 接受日期: 2019-01-21

基金: 国家自然科学基金项目(31572207), 国家科技支撑计划课题(2015BAD23B02), 国家重点研发计划子课题(2018YED0201303)资助

摘要

为了明确不同生育时期进行玉米氮素营养诊断的叶片层位, 建立准确稳健的玉米氮素营养诊断模型, 以达到合理追施氮肥, 提高氮肥利用率的目的。试验采用单因素盆栽试验设计, 以玉米(郑单958)为研究对象, 应用高光谱技术, 分析了不同氮营养水平下不同生育时期不同层位玉米叶片的氮含量分布和变化规律及光谱响应特征; 并依据叶片氮含量与光谱反射率的相关关系, 叶片氮含量与全波段(400~2 000 nm)任意两两波段组合构建的比值光谱指数(RSI)的回归关系, 初步确定了不同生育时期进行氮素营养高光谱诊断的目标叶片, 筛选出最优的比值光谱指数, 建立了叶片氮素含量估算模型。结果表明: 玉米叶片氮含量: 上层>中层>下层; 随着玉米的生长, 在低氮条件下上层叶片氮含量呈先减少后增加(追肥)再减少趋势, 在高氮条件下呈减少趋势, 中下层叶片氮含量呈递减趋势。六叶期下层玉米叶片光谱反射率敏感范围较大, 相关性较强; 九叶期和灌浆期上层玉米叶片的光谱反射率敏感范围较广, 相关性较强; 开花吐丝期中层叶片的光谱反射率敏感范围较大, 相关性较强。六叶期选取下层叶作为诊断目标叶, 选取最佳比值光谱指数RSI(1 811, 1 842)建立线性估算模型, 九叶期和灌浆期选取上层叶片作为诊断目标叶, 选取的最佳比值光谱指数分别为RSI(720, 557), RSI(600, 511)建立线性估算模型, 开花吐丝期选取中层叶片作为诊断目标叶, 选取比值光谱指数RSI(688, 644)建立线性估算模型。研究结果可为快速准确地利用光谱技术进行玉米叶片氮素营养诊断提供理论依据。

关键词: 玉米叶片; 氮素含量; 光谱指数; 分层诊断

中图分类号:O433.4 文献标志码:A

Nitrogen Nutrition Diagnostic Based on Hyperspectral Analysis about Different Layers Leaves in Maize

ZHANG Yin-jie, WANG Lei^*, BAI You-lu, YANG Li-ping, LU Yan-li, ZHANG Jing-jing, LI Ge

Key Laboratory of Plant Nutrition and Fertilizer, Ministry of Agriculture and Rural Affairs, Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China

*Corresponding author

Abstract

In order to clarify the location of nitrogen nutrient diagnosis of maize leaves at different growth stages, and to establish an accurate and robust model to diagnose maize’s nitrogen nutrition, which aims to guide rational fertilization and improve recovery rate, in this experiment, a single factor pot experiment was designed, and maize (Zhengdan 958) was used as the research object to study the distribution and variation of nitrogen content in different layers of leaf under different nitrogen nutrition levels. The distribution and variation of N content and the spectral response characteristics of maize leaves were analyzed. And the correlation relationship between nitrogen content and spectral reflectance of different layers leaves at different growth stages was investigated. Moreover, the regression relationship between the leaf nitrogen content and the ratio spectral index (RSI) which was composed of any two bands between 400~2 000 nm was explored. According to these analyses, leaf layer, optimal RSI and estimation models were initially determined at different stages for nitrogen nutrition diagnosis by spectral technique. The main results are as follows: The results indicate that the maize’s nitrogen content in different layers is as follows: the upper layer>the middle layer>the lower layer; and that as the stages of growth forward, leaves’ nitrogen content in upper layer, under the condition of low-nitrogen, appears to first decrease and then increase (after manuring) and decrease again while keeping the tendency of decrease under the condition of high-nitrogen, with the leaves’ nitrogen content in the levels of middle and low appearing to decrease. At the Six-leaf stage, the lower layer of leaves has a larger sensitivity range and a stronger correlation coefficient. At Nine-leaf and Filling stage, spectral reflectance of the upper layer maize leaves was more sensitive and correlated. At the flowering and silking stage, spectral reflectance of the middle layer leaves was more sensitive and relevant. SO the lower leaves were selected as the diagnosis target at the Six-leaf stage, and the optimal ratio spectral index RSI (1 811, 1 842) was selected to establish the linear estimation model. The upper leaves were selected as the diagnostic target at the Nine-leaf stage and the Filling stage, and the optimal ratio spectral indices were RSI (720, 557), RSI (600, 511) to establish the linear estimation model, respectively. The middle leaves were selected as the diagnostic target during the anthesis-silking stage, and the RSI (688, 644) spectral index was selected to establish the estimation model. The research results could provide a theoretical basis for rapid and accurate nitrogen nutrition spectrum diagnosis method in maize or other crop.

Keyword: Maize leaf; Nitrogen Content; Spectral index; Diagnosis layer

文章图片

引言

氮素是作物生长发育周期中重要的营养元素之一, 及时、准确地进行氮素管理, 既能保证产量又提高肥料效益, 减少环境污染。光谱技术能够快速、便捷、无损地获取作物的相关信息, 在作物氮素光谱营养诊断方面已经开展了许多研究^{[1, 2]}。作物的氮素水平可以通过光谱指数进行反演, Inoue^[3]探究比较了构建的归一化光谱指数(NDSI)、比值光谱指数(RSI)、植被光谱指数以及多元回归统计方法反演水稻氮素水平的能力, 发现构建的比值光谱指数RSI(D740, D522)对氮素的评估是最准确和稳健的。廖钦洪等^[4]比较了15个植被指数, 结果表明构建的归一化光谱指数NDSI (754, 713)相比于其他植被指数更适合生姜叶片氮含量的估算, 模型估测精度R²达0.83。 Wang等^[5]结果表明, RSI(822, 738)光谱指数可以作为监测水稻和小麦叶片氮积累的指标。有研究表明大喇叭口期光谱指数DSI(564, 681)和DSI(681, 707)构建的指数预测模型可以有效估测玉米叶片的氮素含量。

目前已有较多研究结果都显示, 使用特定的光谱参数^[6]和光谱指数^{[7, 8]}可以较准确地预测叶片氮含量, 但不同生育期不同层位的玉米叶片, 光谱响应存在着较大的差异。研究不同层位叶片的光谱特性, 有利于提高光谱诊断的精度, 打破作物种间差异、种植区域和气候条件的限制。本试验根据不同生育期的特点进行了玉米叶片明确的分层, 系统地分析了各生育时期不同层位玉米叶片的光谱响应特征, 选取了各生育时期最适合作为玉米氮素营养诊断的叶片层位及最佳光谱指数, 并建立玉米叶片氮素含量的估算模型, 旨在为及时准确的作物氮素营养诊断及合理施肥提供参考。

1 实验部分

1.1 材料

试验于2017年6月至9月在中国农业科学院国际高新产业园内(廊坊万庄), 供试土壤取自园区试验田, 土壤类型为潮土, 其肥力状况为: 有机质含量2.04 g· kg^-1, pH 8.83, 铵态氮含量13.39 mg· kg^-1, 硝态氮含量14.65 mg· kg^-1。每个处理12盆, 盆直径33 cm, 高40 cm, 每盆装土16 kg, 盆底封闭, 埋于土壤中。

1.2 试验设计

试验供试土壤的常规推荐施肥量为: N: 180 kg· hm^-2, P₂O₅: 90 kg· hm^-2, K₂O : 90 kg· hm^-2(土壤养分系统分析法ASI法^[9]), 依据推荐施肥量分别增施和减施氮肥, 设置6个施氮水平: N0, N1, N2, N3, N4, N5, 施肥量是大田试验的三倍分别为: 0, 180, 360, 540, 720, 900 kg· hm^-2。磷钾肥也按推荐施肥量的3倍施用, 施用量均为270 kg· hm^-2。根据土壤容量(1.13 g· cm^-3)换算成每公斤风干土的施肥量, 土壤与基肥混匀装盆, 施氮肥(尿素, N46%)基追比4:6(8月18日追施尿素), 磷钾肥做基肥一次施入。

表1 采样时间与部位 Table1 Sampling time and location

1.3 样品采集

由表1可知: 六叶期和九叶期的叶片从上往下第一片展开叶开始分层, 开花吐丝期以穗位叶以上三片叶为上层叶, 穗位叶及以下两片叶为中层叶, 中层叶下面三片叶为下层叶片, 灌浆期下层叶片有部分干枯未取样。九叶期与开花吐丝期N0处理下层叶片有部分干枯未取样。各时期测定光谱数据后采集叶片样品进行养分测定。

1.4 高光谱测定

高光谱数据采用FieldSpec光谱仪(美国 ASD公司产)获得, 利用叶片夹持器(自带光源)选取叶片的中部较宽(避免直接对着叶脉)的部分进行光谱测定, 叶片夹持器测定使叶片处于密闭条件下, 手柄中的光源部分直接夹紧玉米叶片所测部位, 保证叶片水平且所测面积相同, 使光谱数据获取的精确性和稳定性提高^[10]。

1.5 数据处理和统计分析

利用光谱处理软件ViewSpecPro5.7导出光谱数据, Excel软件和SPSS22软件进行数据统计分析。借鉴Inoue^[3]等的方法, 本试验考察了全波段范围内任意波段两两组合的比值光谱指数对玉米叶片氮含量的估算效果, 采用波段组合算法, 以矩阵联立的形式, 构建比值光谱指数(ratio spectral index, RSI)公式形式为

$RSI (i, j) = R_{i} / R_{j}$

式中: i和j为400~2 000 nm范围内的任意两个波段, R_i和R_j为波长i和波长j对应的光谱反射率为进一步优化比值光谱指数, 筛选出对叶片氮含量敏感的波段位置, 利用MATLAB编写程序把所有波段两两组合构建的比值光谱指数RSI(i, j)与叶片氮含量线性拟合的决定系数结果用等值线图表示, 颜色的深浅代表了决定系数的大小。

2 结果与讨论

2.1 不同生育时期不同层位叶片氮含量的变化

由表2可知: 六叶期上层叶片氮含量高于下层叶片, 施氮处理叶片氮含量显著高于不施氮处理, 施氮处理之间变化不大, 后三个时期玉米叶片氮含量: 上层> 中层> 下层, 低氮处理间叶片氮含量显著增加, 高氮处理间差异不显著。

在整个玉米生育期中, 六叶期叶片氮含量最高, 随着玉米的生长, 上层叶片低氮处理的氮含量呈先减少后增加(8月18日追肥)再减少趋势, 高氮处理的氮含量呈减少趋势, 中层和下层叶片氮含量呈减少趋势, 这是由于玉米生长中后期营养生长转向生殖生长, 大部分氮素由营养器官转移至生殖器官。在同一生育期, 同一施肥水平下叶片氮含量: 上层> 中层> 下层, 这是由于氮素是易于转运的营养元素。同一生育期同一层位在一定范围内随着氮肥的增施叶片氮含量呈先增加后平稳的趋势。

表2 叶片氮含量(%) Table 2 Leaf nitrogen content(%)

2.2 不同施氮水平下玉米叶片氮含量与光谱反射率的相关分析

由图1(a)可知, 六叶期上层叶片氮含量与1 001~1 133和1 873~1 883 nm波段处光谱反射率呈显著相关关系, 下层叶片氮含量与760~1 147, 1 148~1 163, 1 299~1 358 nm波段处光谱反射率呈极显著相关关系, 六叶期处于苗期与拔节期之间, 而夏玉米在此时期吸收氮素少, 吸收速度慢, 缺氮时下层叶片氮素营养运输到上层叶片, 下层叶片对于氮素缺乏较敏感。由图1(b)可知, 九叶期上层叶片含氮量与411~1 300和1 880~2 001 nm波段处光谱反射率呈极显著相关关系, 中层叶片含氮量与495~735 nm波段处光谱反射率呈极显著相关关系, 下层叶片氮含量与531~718 nm波段处光谱反射率呈极显著相关关系。相比之下, 上层叶片氮含量的光谱反射率的相关性较强, 九叶期处于大喇叭口期左右, 是玉米吸收养分的关键时期, 若氮素得不到及时供应, 上层叶对于氮素的缺乏在光谱的响应较下层叶更为敏感。由图1(c)可知, 开花吐丝期, 在460~729和751~1 296 nm处上层叶片氮含量与光谱反射率呈极显著相关关系, 在474~728, 741~1 300, 1 301~1 391和1 601~1 839 nm处中层叶片氮含量与光谱反射率呈极显著相关关系, 在501~734和1 986~2 047 nm处的光谱反射率与下层叶片氮含量呈极显著相关关系, 相比之下中层叶片氮含量与光谱反射率相关性较强。玉米的开花吐丝期是雌蕊和雄穗形成和发育的时期, 吸收养分氮素营养速度快, 数量多, 其他叶片的氮素运输给上层叶和穗位叶(中层叶), 而穗位叶的氮素又向雌蕊运输, 中层叶片对于氮素的缺乏在光谱响应上较敏感。由图1(d)可知, 灌浆期上层叶片氮含量与508~724, 740~1 300和1 301~2 259 nm处光谱反射率呈极显著相关关系, 中层叶片氮含量与532~718, 736~1 300和1 301~2 229 nm处光谱反射率呈极显著相关关系玉米生长后期(灌浆期), 玉米氮素从营养器官转移到生殖器官中, 其他叶片(上层和下层叶)的氮素大量向果穗叶(中层叶)转移, 穗位叶的氮素再向果穗转移, 上层叶片对于氮素缺乏在光谱响应上较敏感。

	Figure Option View Download New Window
	图1 不同层位玉米叶片氮含量与光谱反射率的相关曲线 (a), (b), (c), (d)代表六叶期, 九叶期, 开花吐丝期, 灌浆期4个生育时期Fig.1 Correlation curvebetween leaf total N and spectral reflectance at different layers (a), (b), (c), (d) represent six-leaf, nine-leaf, anthesis-silking, filling stage four birth stages

2.3 不同生育期不同层位叶片氮含量与比值光谱指数的回归分析

如图2(a)所示, 上层叶片氮含量与比值光谱指数的敏感范围较小, 敏感波段在760 nm处“ 红边位置” 附近, 图2(b)可知, 而下层叶片氮含量估算敏感的范围为760~1 900 nm波段, 其中颜色较深的为1 800 nm附近“ 水分吸收带” 。相比之下, 下层叶片的光谱敏感范围较广, 决定系数也较大, 说明下层叶片的比值光谱指数能更准确地反演氮含量。

	Figure Option View Download New Window
	图2 六叶期叶片氮含量与任意两波段组合RSI(i, j)之间线性拟合的决定系数(R²)的等值线图 (a), (b)分别代表上, 下层位Fig.2 A contour map of the coefficient of determination (R²) between leaf nitrogen content and RSI (i, j) using the thorough combinations of two wavebands at i and j nm (a), (b) represent upper, middle, lower layers

由图3(a, b, c)分别可以看出, 对上层叶片氮含量估算敏感的波段范围较大410~780 nm, 且颜色较深(红), 对中层叶片氮含量估算敏感的波段500~600和700~760 nm, 颜色较上层叶片的浅, 对下层叶片氮含量估算敏感的波段500~580和700~750 nm。九叶期的上中下层叶片氮含量与比值光谱指数敏感波段区域相似, 但上层叶片的光谱敏感范围较广, 决定系数较大, 中层叶片次之, 下层叶片最小, 说明上层叶片比值光谱指数能更准确地反演叶片氮含量。

	Figure Option View Download New Window
	图3 九叶期叶片氮含量与任意两波段组合RSI(i, j)之间线性拟合的决定系数(R²)的等值线图 (a), (b), (c)分别代表上, 中, 下层位Fig.3 A contour map of the coefficient of eetermination (R²) between leaf nitrogen content and RSI(i, j) using the thorough combinations of two wavebands at i and j nm (a), (b), (c) represent upper, middle, lower layers

由图4(a, b, c)分别可知: 对上层叶片氮含量估算敏感波段为500~780 nm, 对中层叶片估算敏感波段范围较大400~790 nm, 且颜色较深(红), 对下层叶片氮含量估算敏感的波段为500~600和750~780 nm。开花吐丝期的各层位叶片氮含量与比值光谱指数敏感波段范围相似, 但中层叶片的敏感范围较大, 且回归方程的决定系数较大, 上层叶片次之, 下层叶片最小, 说明中层叶片比值光谱指数能更准确地反演叶片氮含量。

	Figure Option View Download New Window
	图4 开花吐丝期叶片氮含量与任意两波段组合RSI(i, j)之间线性拟合的决定系数(R²)的等值线图 (a), (b), (c)分别代表上, 中, 下层位Fig.4 A contour map of the coefficient of determination (R²) between leaf nitrogen content and RSI(i, j) using the thorough combinations of two wavebands at i and j nm (a), (b), (c) represent upper, middle, lower layers

由图5(a, b)中可以看出, 对上层叶片氮含量估算敏感波段500~780 nm, 颜色较深(红); 对中层叶片氮含量估算敏感波段500~780 nm, 颜色较浅(黄)。灌浆期的上中层叶片氮含量与比值光谱指数敏感波段一致, 但上层叶片氮含量与比值光谱指数线性回归方程的决定系数较大。因此, 该时期上层叶片的光谱响应更能反映玉米后期的氮素营养状况。

	Figure Option View Download New Window
	图5 灌浆期叶片氮含量与任意两波段组合RSI(i, j)之间线性拟合的决定系数(R²)的等值线图 (a): 上层位; (b): 中层位Fig.5 A contour map of the coefficient of determination (R²) between leaf nitrogen content and RSI(i, j) using the thorough combinations of two wavebands at i and j nm (a): Upper layers; (b): Middle layers

2.4 不同生育期最佳叶片光谱指数与叶片氮含量的回归方程

综合分析, 选取各时期不同处理叶片氮含量与原始光谱反射率及比值光谱指数敏感性较强的叶片作为诊断目标叶, 六叶期选取下层叶片作为叶片氮素诊断的目标叶, 如图6选出最佳的比值光谱指数(ratio spectral index, RSI)(1 811, 1 842)与叶片氮含量建立线性回归方程, 拟合度R²为0.614达极显著水平。九叶期和灌浆期选取上层叶片作为氮素营养诊断的目标叶, 选出最佳比值光谱指数分别为RSI(720, 557), RSI(600, 511)建立线性回归方程, 拟合度R²分别为0.906, 0.889达极显著水平。开花吐丝期选取中层叶片作为叶片氮素营养诊断的目标叶, 选出RSI(688, 644), 建立回归方程, 拟合度R²为0.892达极显著水平。不同时期选取了不同层位的叶片作为光谱诊断的目标叶, 选取的最佳光谱指数的波段组合也不同, 分析认为作物的光谱反射或发射特性是由其化学和形态学特征决定的, 不同层位叶片内部结构存在差异, 作物的形态特征和体内生化组分随着作物生长而不断变化, 导致了不同生育时期不同层位叶片对光谱的响应存在差异。

	Figure Option View Download New Window
	图6 不同生育期最佳比值光谱指数叶片氮素估算回归方程 (a), (b), (c), (d)代表六叶期, 九叶期, 开花吐丝期, 灌浆期4个生育时期Fig.6 Regression equations of leaf nitrogen content and optimal spectral indices in different stag (a), (b), (c), (d) represent six-leaf, nine-leaf, anthesis-silking, filling stage four birth stages

3 结论

分析了夏玉米不同施氮水平下各生育期不同层位叶片氮素含量与光谱反射率的关系, 确定了进行玉米氮素高光谱诊断的叶片层位, 并选取最佳比值光谱指数拟合了各个时期叶片氮含量的估测模型, 得到了以下主要结论:

(1)随着生育期延长, 叶片氮含量呈减少趋势, 在同一生育期, 同一施肥水平下叶片氮含量: 上层> 中层> 下层。在一定范围内同一时期同一层位叶片氮含量随着氮肥的增施呈先增加后平稳趋势。

(2)六叶期下层叶片氮含量与光谱反射率相关关系达显著水平的波段范围更广, 相关性更强。九叶期和灌浆期上层叶片氮含量与光谱反射率相关关系达极显著水平的波段范围更广, 相关性更强。开花吐丝期中层叶片氮含量与光谱反射率相关关系达极显著水平的波段范围更广, 相关性更强。

(3)六叶期选取下层叶片, 九叶期和灌浆期选取上层叶片, 开花吐丝期选取中层叶片作为叶片氮素诊断对象, 分别以RSI(1 811, 1 842), RSI(720, 557), RSI(600, 511)和RSI(688, 644)构建的估算模型反演叶片氮含量效果较好, 其估算模型的精度和稳定性, 以及在不同品种不同地区的应用上仍需要进一步验证。

参考文献

文献列表

[1]	GUO Chao-fan, DUAN Fu-zhou, GUO Xiao-yu, et al(郭超凡, 段福洲, 郭逍宇, 等). Acta Ecologica Sinica(生态学报), 2014, 34(17): 4839. [本文引用:1]
[2]	WANG Ren-hong, SONG Xiao-yu, LI Zhen-hai, et al(王仁红, 宋晓宇, 李振海, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2014, (19): 191. [本文引用:1]
[3]	Inoue Y, Sakaiya E, Zhu Y, et al. Remote Sensing of Environment, 2012, 126(6829): 210. [本文引用:2]
[4]	LIAO Qin-hong, LI Hui-he, ZHANG Qin, et al(廖钦洪, 李会合, 张琴, 等). Research of Agricultural Modernization(农业现代化研究), 2017, 38(2): 315. [本文引用:1]
[5]	Wang Wei, Yao Xia, Tian Yongchao, et al. Journal of Integrative Agriculture, 2012, 11(12): 2001. [本文引用:1]
[6]	LI Yuan-yuan, CHANG Qing-rui, LIU Xiu-ying, et al(李媛媛, 常庆瑞, 刘秀英, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2016, 32(16): 135. [本文引用:1]
[7]	ZHANG Xiao-yuan, ZHANG Li-fu, ZHANG Xia, et al(张潇元, 张立福, 张霞, 等). Scientia Agricultura Sinica(中国农业科学), 2017, 50(3): 474. [本文引用:1]
[8]	Jin Xiuliang, Xu Xingang, Feng Haikun, et al. International Journal of Agriculture and Biology, 2014, 16(3): 498. [本文引用:1]
[9]	JIN Ji-yun, BAI You-lu, YANG Li-ping(金继运, 白由路, 杨俐苹). High Efficient Technology and Equipment for Soil Testing( 高效土壤养分测试技术与设备). Beijing: China Agriculture Press(北京: 中国农业出版社), 2006. 148. [本文引用:1]
[10]	LIU Xiao-jun, TIAN Yong-chao, YAO Xia, et al(刘小军, 田永超, 姚霞, 等). Scientia Agricultura Sinica(中国农业科学), 2012, 45(3): 435. [本文引用:1]

2014

0.0

GUO

Chao-fan

, DUAN

Fu-zhou

, GUO

Xiao-yu

, et al(郭超凡, 段福洲, 郭逍宇, 等). Acta Ecologica Sinica(生态学报), 2014, 34(17): 4839.

Utilization of hyperspectral remote sensing technology to estimate wetland plant leaf nitrogen content quantitatively in large area is important to monitor and diagnosis of physiological condition and growth trend of wetland vegetation. However, there are many limitations such as over-fitting of inversion model, indeterminate causal relationship between the selected bands and biochemical parameters,"multicollinearity" of selected bands in leaf nitrogen diagnosis remote sensing research. The total nitrogen content of leaves of typical wetland plants, Phragmites australis and Typha angustifolia, was selected as our study objects. These plants grow in South Wetland purification system in the Olympic Park in Beijing, a typical wetland using reclaimed water. The leaf reflectance spectra of main wetland plants were acquired by means of an ASD FieldSpec 3 spectrometer (350-2500nm). Leaf total nitrogen content was determined by Kjeldahl nitrogen measurement method after acquiring the leaf reflectance spectra. The method,"Give a cross correlation analysis", was used to build the correlations between leaf nitrogen content and the original spectrum, first derivative spectrum. Then, the selected bands were divided into some areas according to QiSeGuang spectral range and the interval of selected bands. Those bands which are high frequency、frequency more、ranked high frequency are chosen as representative bands of different areas and considered to be the optimal bands used to build the regression model. Finally,partial least squares was used to build inversion model. The accuracy of this model was tested with cross-validated coefficient of determination ( R cv 2 ) and cross、|validated root mean square error (RMSE cv ).The results show that the first derivative transformation can effectively improve the sensitivity of the original spectrum leaf nitrogen content inversion, and fully reflect the sensitivity of near infrared wave band representing leaf total nitrogen content. The accuracy of regression model based on the first derivative and the partial least-squares was much higher than that of the original spectra. In the regression model of the reed, verification accuracy ( R cv 2 ) reached 0.84, square root error (RMSE cv ) was 0.11, in the regression model of the cattail, verification accuracy ( R cv 2 ) reached 0.66, square root error (RMSE cv ) was 0.13, which were the optimal models to estimate leaf total nitrogen content. The determination of parameters in "Give a cross correlation analysis" provided a scientific basis for building the model of eliminating "singularband" and reducing "multicollinearity" problem. Spectrum zone division provides a scientific basis for revealing the causal relationships between optimal band and biochemical parameters. And partial least squares regression method was used to avoid "multicollinearity" of selected bands. The result from this study can not only fill the gaps in the detection of leaf nitrogen using remote sensing, but also provide a strong scientific basis for the nitrogen content monitoring and management of urban wetlands using reclaimed water. At last, Partial least squares regression method was used to avoid "multicollinearity" of selected bands.

利用高光谱遥感技术定量估测湿地植被叶片全氮含量，对于监测和诊断湿地植被的生理状况及生长趋势具有重要意义。但叶片氮素遥感诊断研究多存在反演模型过拟合、入选波段与生化参量间因果关系不明确和入选变量间"多重共线性"等局限。以芦苇（ Phragmites australis ）和香蒲（ Typha angustifolia ）叶片全氮含量作为研究对象，通过谱带分区，分区最佳波段选取和偏最小二乘回归相结合的方法构建芦苇和香蒲叶片全氮含量反演模型，并利用交叉验证决定系数（ R cv 2 ）和均方根误差（RMSE cv ）对模型精度进行检验，尝试克服传统反演方法中的不足。结果表明，不同湿地植物类型相比，利用芦苇反射光谱建立的预测模型精度都高于香蒲。不同回归模型相比，一阶导数光谱-偏最小二乘回归模型（FDS-PLSR）精度远高于原始光谱-偏最小二乘回归模型（OS-PLSR）。芦苇最佳模型交叉验证决定系数（ R cv 2 ）达到了0.84，方根误差（RMSE cv ）为0.10，香蒲最佳模型交叉验证的决定系数（ R cv 2 ）达到了0.66，方根误差（RMSE cv ）为0.13，是构建湿地植物芦苇和香蒲光谱与叶片全氮含量关系的最佳模型。在不降低湿地植物叶片氮含量反演精度的基础上，有效地避免了传统地物高光谱模型反演中的局限性，是无损害遥感探测方面的有益尝试。

... 光谱技术能够快速、便捷、无损地获取作物的相关信息, 在作物氮素光谱营养诊断方面已经开展了许多研究^[1,2] ...

2014

0.0

WANG

Ren-hong

, SONG

Xiao-yu

, LI

Zhen-hai

, et al(王仁红, 宋晓宇, 李振海, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2014, (19): 191.

为了准确定量诊断氮素状况，为施肥和产量、品质的估测提供参考，该文通过设置不同氮素水平和品种类型的冬小麦田间试验，分析孕穗至灌浆初期不同光谱参数在小麦氮素营养状况监测上的差异，筛选叶片氮素含量和冠层氮素密度反演效果较好的参数，建立其与氮营养指数（NNI，nitrogen nutrition index）的经验模型。研究表明，线性内插法红边位置（REPLI）、修正红边单比指数（mSR705）、比值指数（RI-1dB）、简单比值色素指数（SRPI）、红边指数（VOG）等光谱参数与氮素营养指标具有良好的相关性(r>0.85)，且不受生育期影响，可用来反演评价冠层氮素营养状况；研究对筛选的光谱参数与各氮素指标进行回归建模，并用独立试验数据对所建模型进行验证，结果显示，REPLI在氮营养指数估测方面表现较好(r=0.93)，估测模型精度较高（决定系数R2=0.86，均方根误差RMSE = 0.08）。NNI在氮素营养状况诊断方面有一定的优势，通过高光谱反演氮营养指数进行氮素营养状态的定性定量诊断有一定的可行性。

... 光谱技术能够快速、便捷、无损地获取作物的相关信息, 在作物氮素光谱营养诊断方面已经开展了许多研究^[1,2] ...

2012

0.0

... 作物的氮素水平可以通过光谱指数进行反演, Inoue^[3]探究比较了构建的归一化光谱指数(NDSI)、比值光谱指数(RSI)、植被光谱指数以及多元回归统计方法反演水稻氮素水平的能力, 发现构建的比值光谱指数RSI(D740, D522)对氮素的评估是最准确和稳健的 ...

... 借鉴Inoue^[3]等的方法, 本试验考察了全波段范围内任意波段两两组合的比值光谱指数对玉米叶片氮含量的估算效果, 采用波段组合算法, 以矩阵联立的形式, 构建比值光谱指数(ratio spectral index, RSI)公式形式为 ...

2017

0.0

... 廖钦洪等^[4]比较了15个植被指数, 结果表明构建的归一化光谱指数NDSI (754, 713)相比于其他植被指数更适合生姜叶片氮含量的估算, 模型估测精度R²达0 ...

2012

0.0

... Wang等^[5]结果表明, RSI(822, 738)光谱指数可以作为监测水稻和小麦叶片氮积累的指标 ...

2016

0.0

Yuan-yuan

, CHANG

Qing-rui

, LIU

Xiu-ying

, et al(李媛媛, 常庆瑞, 刘秀英, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2016, 32(16): 135.

为了进一步提高玉米叶绿素含量的高光谱估算精度，该文测定了西北地区玉米乳熟期叶片的光谱反射率及其对应的叶绿素相对含量（soil and plant analyzer development，SPAD）值，分析了一阶微分光谱、高光谱特征参数与SPAD的相关关系，构建了基于一阶微分光谱、高光谱特征参数和BP神经网络的SPAD估算模型，并对模型进行验证；再结合主成分回归（principal component regression，PCR）、偏最小二乘回归（partial least squares regression，PLSR）以及传统回归模型与BP神经网络模型进行比较。结果表明：SPAD值与一阶微分光谱在763nm处具有最大相关系数（R=0.901）；以763 nm处的一阶微分值、蓝边内最大一阶微分为自变量建立的传统回归模型可用于玉米叶片SPAD估算；将构建传统回归模型时筛选到的光谱参数作为输入，实测SPAD值作为输出，构建BP神经网络模型，其建模与验模R2分别为0.887和0.896，RMSE为2.782，RE为4.59%，与其他回归模型相比，BP神经网络模型预测精度最高，研究表明BP神经网络对叶绿素具有较好的预测能力，是估算玉米叶片SPAD值的一种实时高效的方法。

... 目前已有较多研究结果都显示, 使用特定的光谱参数^[6]和光谱指数^[7,8]可以较准确地预测叶片氮含量, 但不同生育期不同层位的玉米叶片, 光谱响应存在着较大的差异 ...

2017

0.0

ZHANG

Xiao-yuan

, ZHANG

Li-fu

, ZHANG

Xia

, et al(张潇元, 张立福, 张霞, 等). Scientia Agricultura Sinica(中国农业科学), 2017, 50(3): 474.

【Objective】Nitrogen is one of the most important nutrients in crop growth and development. The objective of this paper is to study the setting of effective index of leaf nitrogen content inversion in order to provide an important basis for the application of hyperspectral vegetation index of leaf nitrogen content estimation, and for real-time monitoring and accurate diagnosis of crops.【Method】 A total of 225 groups of canopy reflectance and leaf nitrogen content data which covering the whole winter wheat growth period and under different levels of coverage, were collected to simulate different spectral index like different central wavelengths, SNR and band width indicators, and to analyze the influence of different observation pattern on quantitative models. And then, the indicators of accuracy evaluation, coefficient of determination, root mean square error, mean absolute error, mean relative error and P 0.01 were used to select the optimal model and the best indicators, and the sensitivity and effectiveness of leaf nitrogen content quantitative models inversion were analyzed with different spectral indicators. 【Result】MTCI_B was the best vegetation index for leaf nitrogen content inversion with the center wavelengths of 420 nm, 508 nm and 405 nm, band width of 1nm, SNR greater than 70 DB; the correlation with measured nitrogen content was preferably ( R 2 =0.7674, RMSE =0.5511% , MAE =0.4625%, MRE =11.11 percentage points and P inf_r was the best index for inversion of high coverage with the optimal center wavelengths 826 nm and 760 nm ( R 2 =0.6739, RMSE =0.2964%, MAE =0.2851%, MRE =6.44 percentage points and P R 2 =0.8252, RMSE =0.4032%, MAE =0.4408%, MRE =12.22 percentage points and P inf_r and MTCI to build a joint inversion model, the model accuracy evaluation result ( R ²=0.9286, RMSE =0.3416%, MAE =0.2988%, MRE =7.16 percentage points and P R 2 reached 0.92-0.93, RMSE were between 0.37%-0.39%) was better than the single vegetation index ( R 2 were 0.79-0.81, RMSE were between 0.63%-0.66%).【Conclusion】A good estimation of crop leaf nitrogen content could be realized by using hyperspectral vegetation index, quantitative inversion of crop leaf nitrogen content had a strong sensitivity with different spectral indexes, center wavelength, SNR, and band width. Application of multi-exponential joint inversion model significantly improved the accuracy of the inversion. And the joint inversion model had a certain degree of universality in different hyperspectral sensors.

1 Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101 ; 2 School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083

【目的】氮素是作物生长发育过程中最重要的营养元素之一,研究叶氮含量反演的有效光谱指标设置,为应用高光谱植被指数反演作物叶氮含量,以及作物的实时监测与精确诊断提供重要依据。【方法】以冬小麦为例,选取涵盖冬小麦全生育期不同覆盖程度 225组冠层光谱与叶氮含量数据,通过遥感方法建立模型,模拟了不同光谱指标,即中心波长、信噪比和波段宽度对定量模型的影响,通过模型精度评价指标决定系数（coefficient of determination, R 2 ）、根均方差（root mean square error, RMSE ）、平均绝对误差（mean absolute error, MAE ）、平均相对误差（ mean relative error, MRE ）和显著性检验水平（ P R 2 =0.7674, RMSE =0.5511%, MAE =0.4625%, MRE =11.11个百分点,且 P inf_r （ R 2 =0.6739, RMSE =0.2964%, MAE =0.2851%, MRE =6.44个百分点,且 P R 2 =0.8252, RMSE =0.4032%, MAE =0.4408%, MRE =12.22个百分点,且 P inf_r 和MTCI构建的联合反演模型（ R 2 =0.9286, RMSE =0.3416%, MAE =0.2988%, MRE =7.16个百分点,且 P R 2 为0.92—0.93, RMSE 在0.37%—0.39%, MAE 为0.285%左右, MRE 约为7.00个百分点）明显优于单一植被指数反演精度（ R 2 为0.79—0.81, RMSE 为0.63%—0.66%, MAE 为0.455%左右, MRE 约为10.90个百分点）。【结论】利用高光谱植被指数可有效实现作物叶氮含量反演,作物叶氮含量定量反演对不同光谱指标 —中心波长、信噪比和波段宽度,具有较强敏感性。应用多指数联合反演模型,可显著提高反演精度,并且联合反演模型在不同高光谱传感器下有一定普适性。

2014

0.0

2006

0.0

... hm^-2(土壤养分系统分析法ASI法^[9]), 依据推荐施肥量分别增施和减施氮肥, 设置6个施氮水平: N0, N1, N2, N3, N4, N5, 施肥量是大田试验的三倍分别为: 0, 180, 360, 540, 720, 900 kg#cod#x000b7 ...

2012

0.0

LIU

Xiao-jun

, TIAN

Yong-chao

, YAO

Xia

, et al(刘小军, 田永超, 姚霞, 等). Scientia Agricultura Sinica(中国农业科学), 2012, 45(3): 435.

【Objective】 The objective of the experiments is to develop a key method for fast and nondestructive monitoring leaf water content (LWC) in rice (Oryza sativa L.). 【Method】 Two field experiments were conducted with different soil water conditions and rice cultivars across two growing seasons, and time-course measurements were taken on leaf hyperspectral reflectance and LWC at top four leaves over main growth stages. Several kinds of hyperspectral indices at leaf scale including ratio spectral indices (RSI), normalized difference spectral indices (NDSI) and difference spectral indices (DSI) with all combinations of two wavebands between 350 and 2 500 nm were calculated, and their relationships to LWC were analyzed. 【Result】 The results indicated that the leaf spectral reflectance varied distinctly with soil water treatments and different top leaves, the sensitivity bands mostly occured within near-infrared and short-infrared spectral regions. The spectral indices as RSI (R1402, R2272) and NDSI (R1402, R2272) were linear with LWC, giving the determination coefficient of linear regression (S-R2) of 0.80, and the predicted R2 (P-R2) based on the testing performance with independent datasets as 0.86. 【Conclusion】 It is concluded that the RSI (R1402, R2272) and NDSI (R1402, R2272) can be used to monitor leaf water content in rice.

1.南京农业大学农学院/国家信息农业工程技术中心/江苏省信息农业高技术研究重点实验室，南京210095

【目的】建立快速、无损诊断水稻叶片含水量的估测模型，为水稻水分精确管理提供依据。【方法】基于2年不同土壤水分处理和水稻品种的池栽试验，于水稻主要生育时期同步测定顶部4张叶片的光谱反射率和含水量，系统分析350—2 500 nm波段范围内任意两波段组合而成的比值（RSI）、归一化差值（NDSI）及差值（DSI）光谱指数，并分析其与叶片含水量的量化关系。【结果】不同土壤水分处理和叶位间，叶片反射光谱具有显著的时空变化特征，叶片含水量的敏感光谱波段主要位于近红外及短波红外区域；RSI （R1402, R2272）及NDSI （R1402, R2272）光谱指数与叶片含水量呈现良好的线性相关，线性拟合R2均达到0.80。基于独立试验资料对所建模型进行测试检验也显示，预测值和观察值的拟合R2也均达到0.86。【结论】RSI（R1402, R2272）、NDSI（R1402, R2272）均可用于水稻叶片含水量的定量监测。

... 4 高光谱测定高光谱数据采用FieldSpec光谱仪(美国 ASD公司产)获得, 利用叶片夹持器(自带光源)选取叶片的中部较宽(避免直接对着叶脉)的部分进行光谱测定, 叶片夹持器测定使叶片处于密闭条件下, 手柄中的光源部分直接夹紧玉米叶片所测部位, 保证叶片水平且所测面积相同, 使光谱数据获取的精确性和稳定性提高^[10] ...