Chlorophyll Content Inversion of Spring Wheat Based on Unmanned Aerial Vehicle Hyperspectral and Integrated Learning

doi:10.13304/j.nykjdb.2024.0010

Abstract

Abstract:

Chlorophyll content is a key indicator for monitoring crop growth， and its rapid， effective and accurate estimation is crucial for assessing crop health. By collecting unmanned aerial vehicle （UAV） hyperspectral images from 3 growth stages and combining them with ground-measured chlorophyll data， various machine learning and ensemble learning models were employed to estimate the chlorophyll content in spring wheat，and the estimation accuracy of different models were compared. The results showed that the canopy reflectance of spring wheat was generally consistent across different growth stages， but significant differences in spectral reflectance intensity were observed in the 770~900 nm wavelength range. 16 spectral indices all showed significant correlations with chlorophyll content， among which optimized vegetation index 1， plant biochemical index and normalized difference red edge index maintained high correlation throughout the entire growth cycle. The prediction accuracy of the Stacking and Voting ensemble learning models was higher than the basic models， with the Voting ensemble model performing particularly well. In the test set， determination coefficient （R²）values of 3 growth stages were 0.78， 0.77 and 0.73， and root mean square error （RMSE） values were 8.70， 11.36 and 16.17， respectively. Compared with random forest， support vector regression， K_nearest neighbor and ridge regression models， the R² of the Voting model was on average 0.17， 0.14 and 0.22 higher， and the RMSE was 4.64， 2.54 and 6.51 lower， indicating its superior predictive ability. Above results provided new perspectives and methods for precision agriculture and crop health monitoring.

Key words: unmanned aerial vehicle（UAV）, hyperspectral remotesensing, spring wheat, chlorophyll content, ensemble learning

摘要：

叶绿素含量是监测作物长势的关键指标，快速、有效、准确的估算对作物健康评估具有重要意义。通过采集3个生长期的无人机高光谱影像，结合地面叶绿素实测数据，选用多种机器学习和集成学习模型，反演春小麦叶绿素含量，并对比不同模型的反演精度。结果表明，春小麦不同生长期冠层反射率基本一致，但在770~900 nm 波长范围内显示出明显的光谱反射率强度差异。16种光谱指数均与叶绿素含量呈显著相关，其中优化植被指数1、植物生化指数和归一化差异红边指数在整个生长周期内均与叶绿素含量保持高相关性。Stacking和Voting集成学习模型的预测精度高于基础模型，其中Voting集成学习模型表现更突出，测试集中3个生长期的决定系数（R²）分别为0.78、0.77和0.73，均方根误差（root mean square error，RMSE）分别为8.70、11.36和16.17；与随机森林、支持向量机、K_近邻和岭回归相比，其R²平均分别提高约0.17、0.14和0.22，RMSE平均降低4.64、2.54和6.51，显示出良好的预测能力。研究结果可为精准农业和作物健康监测提供新的视角和方法。

关键词: 无人机, 高光谱遥感, 春小麦, 叶绿素含量, 集成学习

CLC Number:

S127

Sile HU, Yulong BAO, Tubuxinbayaer, Jifeng TAO, Enliang GUO. Chlorophyll Content Inversion of Spring Wheat Based on Unmanned Aerial Vehicle Hyperspectral and Integrated Learning[J]. Journal of Agricultural Science and Technology, 2025, 27(6): 93-103.

呼斯乐, 包玉龙, 图布新巴雅尔null, 陶际峰, 郭恩亮. 基于无人机高光谱和集成学习的春小麦叶绿素含量反演[J]. 中国农业科技导报, 2025, 27(6): 93-103.

Figures/Tables 10

Fig. 1 Hyperspectral data splicing process

Fig. 2 Integrated learning flow for LCC estimation

Table 1 Characteristics of changes in LCC in spring wheat at different growth stages

生长期

Growth stage

平均值

Mean value

最大值

Maximum value

最小值

Minimum value

标准差

Standard deviation

变异系数

Coefficient of variation/%

Fig. 3 Box plot of LCC changes in spring wheat at different growth stages

Fig. 4 Spectral reflectance of spring wheat at different growth stages

Fig. 5 Heatmap of the correlation between spectral indices and LCCNote： B—Blue；G—Green；R—Red；NIR—Near infrared；RVI—Ratio vegetation index；NDVI—Normalized difference vegetation index；EVI—Enhanced vegetation index；EVI 2—Two-band enhanced vegetation index；OVG1—Optimized vegetation index 1；NDRE—Normalized difference red edge index；PBI—Plant biochemical index；NDVI705—Normalized difference vegetation index705；mSR705—Modified simple ratio705；PRI—Photochemical reflectance index；SIPI—Structure insensitive pigment index；PSRI—Plant senescence reflectance Index；LCC—Leaf chlorophyll content. * and ** indicate significant correlations at P<0.05 and P<0.01 levels， respectively.

Table 2 Training set for LCCinversion model of spring wheat at different growth stages

生长期 Growth stage	评价指标 Evaluation index	随机森林 RF	支持向量机 SVR	K_近邻 K_NN	岭回归 RR	Stacking	Voting
拔节期 Jointing stage	决定系数R²	0.73	0.66	0.66	0.51	0.77	0.80
拔节期 Jointing stage	均方根误差RMSE	11.94	14.07	14.61	15.78	12.21	11.50
抽穗期 Heading stage	决定系数R²	0.71	0.64	0.61	0.57	0.72	0.80
抽穗期 Heading stage	均方根误差RMSE	12.52	14.50	17.29	15.11	12.25	10.77
灌浆期 Filling stage	决定系数R²	0.71	0.55	0.44	0.43	0.71	0.75
灌浆期 Filling stage	均方根误差RMSE	16.60	22.69	27.92	22.64	17.10	15.52

Table 3 Test set for LCC inversion model of spring wheat at different growth stages

生长期 Growth stage	评价指标 Evaluation index	随机森林 RF	支持向量机 SVR	K_近邻 K_NN	岭回归 RR	Stacking	Voting
拔节期 Jointing stage	决定系数R²	0.72	0.64	0.60	0.49	0.75	0.78
拔节期 Jointing stage	均方根误差RMSE	13.73	10.47	13.10	16.06	9.39	8.70
抽穗期 Heading stage	决定系数R²	0.68	0.61	0.64	0.59	0.70	0.77
抽穗期 Heading stage	均方根误差RMSE	13.24	13.19	14.00	15.20	13.13	11.36
灌浆期 Filling stage	决定系数R²	0.67	0.50	0.42	0.44	0.68	0.73
灌浆期 Filling stage	均方根误差RMSE	17.83	23.67	26.17	23.09	19.98	16.18

Fig. 6 Scatter plots of LCC inversion in spring wheat at different growth stages based on the Voting ensemble learning model

Fig. 7 Inversion graphs of LCC changes in spring wheat at different growth stages

References 45

1	WIJESINGHA J, DAYANANDA S, WACHENDORF M, et al.. Comparison of spaceborne and UAV-borne remote sensing spectral data for estimating monsoon crop vegetation parameters [J/OL]. Sensors,2021,21(8):2886 [2023-11-15]. .
2	BOHMAN B J, ROSEN C J, MULLA D J. Evaluation of variable rate nitrogen and reduced irrigation management for potato production [J]. Agron. J., 2019, 111(4): 2005-2017.
3	LI X Q, LI L, LIU X N. Collaborative inversion heavy metal stress in rice by using two-dimensional spectral feature space based on HJ-1 A HSI and radarsat-2 SAR remote sensing data [J]. Int. J. Appl. Earth Obs. Geoinf., 2019, 78: 39-52.
4	PUTRA B T W.New low-cost portable sensing system integrated with on-the-go fertilizer application system for plantation crops [J/OL]. Measurement, 2020, 155: 107562 [2023-11-15]. .
5	刘涛,张寰,王志业,等.利用无人机多光谱估算小麦叶面积指数和叶绿素含量[J]. 农业工程学报, 2021, 37(19):65-72.
	LIU T, ZHANG H, WANG Z Y, et al.. Estimation of the leaf area index and chlorophyll content of wheat using UAV multi-spectrum images [J]. Trans. Chin. Soc. Agric. Eng., 2021,37(19): 65-72.
6	赵春江.农业遥感研究与应用进展[J].农业机械学报,2014,45(12): 277-293.
	ZHAO C J. Advances of research and application in remote sensing for agriculture [J]. Trans. Chin. Soc. Agric. Mach., 2014,45(12): 277-293.
7	杨贵军,李长春,于海洋,等.农用无人机多传感器遥感辅助小麦育种信息获取[J].农业工程学报, 2015, 31(21): 184-190.
	YANG G J, LI C C, YU H Y, et al.. UAV based multi-load remote sensing technologies for wheat breeding information acquirement [J]. Trans. Chin. Soc. Agric. Eng., 2015, 31(21):184-190.
8	刘建刚,赵春江,杨贵军,等.无人机遥感解析田间作物表型信息研究进展[J].农业工程学报, 2016, 32(24): 98-106.
	LIU J G, ZHAO C J, YANG G J, et al.. Review of field-based phenotyping by unmanned aerial vehicle remote sensing platform [J]. Trans. Chin. Soc. Agric. Eng., 2016, 32(24): 98-106.
9	边明博,马彦鹏,樊意广,等.融合无人机多源传感器的马铃薯叶绿素含量估算[J].农业机械学报,2023,54(8):240-248.
	BIAN M B, MA Y P, FAN Y G, et al.. Estimation of potato chlorophyll content based on UAV multi-source sensor [J]. Trans. Chin. Soc. Agric. Mach., 2023, 54(8): 240-248.
10	QIAO L, TANG W J, GAO D H, et al.. UAV-based chlorophyll content estimation by evaluating vegetation index responses under different crop coverages [J/OL]. Comput. Electron. Agric., 2022, 196:106775 [2023-11-15]. .
11	BEHMANN J, MAHLEIN A K, RUMPF T, et al.. A review of advanced machine learning methods for the detection of biotic stress in precision crop protection [J]. Precis. Agric., 2015,16(3): 239-260.
12	SHU M Y, SHEN M Y, ZUO J Y, et al.. The application of UAV-based hyperspectral imaging to estimate crop traits in maize inbred lines [J/OL]. Plant Phenomics,2021,2021: 9890745 [2023-11-15]. .
13	刘楠,李斐,杨海波,等.优化光谱指数助力机器学习提高马铃薯叶绿素含量反演精度[J].植物营养与肥料学报,2023,29(8):1531-1542.
	LIU N, LI F, YANG H B, et al.. Machine learning models fed with optimized spectral indices to improve inversion accuracy of potato chlorophyll content [J]. J. Plant Nutr. Fert., 2023,29(8):1531-1542.
14	QI H X, WU Z Y, ZHANG L, et al.. Monitoring of peanut leaves chlorophyll content based on drone-based multispectral image feature extraction [J/OL]. Comput. Electron. Agric., 2021,187: 106292 [2023-11-15]. .
15	张卓然,常庆瑞,张廷龙,等.基于支持向量机的棉花冠层叶片叶绿素含量高光谱遥感估算[J].西北农林科技大学学报(自然科学版), 2018, 46(11): 39-45.
	ZHANG Z R, CHANG Q R, ZHANG T L, et al.. Hyperspectral estimation of chlorophyll content of cotton canopy leaves based on support vector machine [J]. J. Northwest A&F Univ. (Nat.Sci.), 2018, 46(11): 39-45.
16	JIAO Q J, SUN Q, ZHANG B, et al.. A random forest algorithm for retrieving canopy chlorophyll content of wheat and soybean trained with PROSAIL simulations using adjusted average leaf angle [J/OL]. Remote. Sens., 2021, 14(1):98 [2023-11-15]. .
17	罗丹,常庆瑞,齐雁冰.基于红边参数和人工神经网络的苹果叶片叶绿素含量估算[J].西北农林科技大学学报(自然科学版),2019,47(1):107-115.
	LUO D, CHANG Q R, QI Y B. Estimation of chlorophyll content in apple leaves based on red edge parameters and artificial neural network [J]. J. Northwest A&F Univ. (Nat. Sci.), 2019, 47(1): 107-115.
18	FEI S P, HASSAN M A, HE Z H, et al.. Assessment of ensemble learning to predict wheat grain yield based on UAV-multispectral reflectance [J/OL]. Remote. Sens., 2021, 13(12): 2338 [2023-11-15]. .
19	ZHANG Z, PASOLLI E, CRAWFORD M M. An adaptive multiview active learning approach for spectral-spatial classification of hyperspectral images [J]. IEEE Trans. Geosci. Remote. Sens., 2020, 58(4):2557-2570.
20	符欣彤,常庆瑞,张佑铭,等.基于Stacking集成学习的猕猴桃叶片叶绿素含量估算[J].干旱地区农业研究,2023,41(4):247-256.
	FU X T, CHANG Q R, ZHANG Y M, et al.. Estimation of kiwifruit leaf chlorophyll content based on Stacking ensemble learning [J]. Agric. Res. Arid Areas, 2023,41(4):247-256.
21	PEPPES N, DASKALAKIS E, ALEXAKIS T, et al.. Performance of machine learning-based multi-model voting ensemble methods for network threat detection in agriculture 4.0 [J]. Sensors, 2021, 21(22): 7475 [2023-11-15]. .
22	LIANG L, DI L P, ZHANG L P, et al.. Estimation of crop LAI using hyperspectral vegetation indices and a hybrid inversion method [J]. Remote. Sens. Environ., 2015, 165:123-134.
23	RABIDEAU S W, LEMONS J F. The potential of the Pu(Ⅲ)—Pu(IV) couple and the equilibrium constants for some complex ions of Pu(IV)¹ [J]. J. Am. Chem. Soc., 1951, 73(6): 2895-2899.
24	DEFRIES R S, TOWNSHEND J R G. NDVI-derived land cover classifications at a global scale [J]. Int. J. Remote. Sens., 1994,15(17): 3567-3586.
25	HUETE A, DIDAN K, LEEUWEN L W, et al.. MODIS vegetation indices [J/OL]. Land Remote Sens. Glob. Environ. Change, 2010:579-602 [2023-11-15]. .
26	JIANG Z Y, HUETE A R, DIDAN K, et al.. Development of a two-band enhanced vegetation index without a blue band [J].Remote. Sens. Environ., 2008, 112(10):3833-3845.
27	VOGELMANN J E, ROCK B N, MOSS D M. Red edge spectral measurements from sugar maple leaves [J]. Int. J. Remote.Sens.,1993, 14(8):1563-1575.
28	FITZGERALD G J, RODRIGUEZ D, CHRISTENSEN L K, et al..Spectral and thermal sensing for nitrogen and water status in rainfed and irrigated wheat environments [J]. Precis. Agric., 2006, 7(4): 233-248.
29	RAO N R, GARG P K, GHOSH S K, et al.. Estimation of leaf total chlorophyll and nitrogen concentrations using hyperspectral satellite imagery [J]. J. Agric. Sci., 2008, 146(1):65-75.
30	WU C Y, NIU Z, TANG Q, et al.. Estimating chlorophyll content from hyperspectral vegetation indices:modeling and validation [J]. Agric. For. Meteor., 2008, 148:1230-1241.
31	DROLET G G, HUEMMRICH K F, HALL F G, et al.. A MODIS-derived photochemical reflectance index to detect inter-annual variations in the photosynthetic light-use efficiency of a boreal deciduous forest [J]. Remote. Sens. Environ., 2005, 98(2): 212-224.
32	SIMS D A, GAMON J A. Relationships between leaf pigment content and spectral reflectance across a wide range of species,leaf structures and developmental stages [J]. Remote. Sens. Environ., 2002, 81(3):337-354.
33	ANDEREGG J, YU K, AASEN H, et al.. Spectral vegetation indices to track senescence dynamics in diverse wheat germplasm [J/OL]. Front. Plant Sci., 2019,10:1749 [2023-11-15]. .
34	张东彦,韩宣宣,林芬芳,等.基于多源无人机影像特征融合的冬小麦LAI估算[J].农业工程学报,2022,38(9):171-179.
	ZHANG D Y, HAN X X, LIN F F, et al.. Estimation of winter wheat leaf area index using multi-source UAV image feature fusion [J]. Trans. Chin. Soc. Agric. Eng., 2022, 38(9):171-179.
35	刘京,常庆瑞,刘淼,等.基于SVR算法的苹果叶片叶绿素含量高光谱反演[J].农业机械学报,2016,47(8):260-265, 272.
	LIU J, CHANG Q R, LIU M, et al.. Chlorophyll content inversion with hyperspectral technology for apple leaves based on support vector regression algorithm [J]. Trans. Chin. Soc. Agric. Mach., 2016, 47(8): 260-265, 272.
36	COSENZA DIOGO N, LAURI K, MATTI M, et al.. Comparison of linear regression, k-nearest neighbour and random forest methods in airborne laser-scanning-based prediction of growing stock [J]. For. Int. J. For. Res., 2021, 94(2):311-323.
37	JI S, GU C, XI X B, et al.. Quantitative monitoring of leaf area index in rice based on hyperspectral feature bands and ridge regression algorithm [J/OL]. Remote Sens., 2022, 14(12): 2777 [2023-11-15]. .
38	杨红云,郭紫微,郭高飞,等.基于Stacking集成卷积神经网络的水稻氮素营养诊断[J].植物营养与肥料学报,2023,29(3):573-581.
	YANG H Y, GUO Z W, GUO G F, et al.. Rice nitrogen nutrition diagnosis based on Stacking integrated convolutional neural network [J]. J. Plant Nutr. Fert., 2023, 29(3):573-581.
39	MAIRE G L, FRANÇOIS C, DUFRÊNE E. Towards universal broad leaf chlorophyll indices using PROSPECT simulated database and hyperspectral reflectance measurements [J]. Remote. Sens. Environ., 2004, 89(1):1-28.
40	汪旭,邓裕帅,练雪萌,等.基于无人机多光谱技术的甜菜冠层叶绿素含量反演[J].中国糖料,2022,44(4):36-42.
	WANG X, DENG Y S, LIAN X M, et al.. Inversion of chlorophyll content in sugar beet canopy based on UAV multispectral technique [J]. Sugar Crops China, 2022, 44(4): 36-42.
41	CHLINGARYAN A, SUKKARIEH S, WHELAN B. Machine learning approaches for crop yield prediction and nitrogen status estimation in precision agriculture:a review [J]. Comput. Electron. Agric., 2018, 151:61-69.
42	LIAKOS K G, BUSATO P, MOSHOU D, et al.. Machine learning in agriculture:a review [J/OL]. Sensors, 2018,18(8):2674 [2023-11-15]. .
43	FENG L W, ZHANG Z, MA Y C, et al.. Alfalfa yield prediction using UAV-based hyperspectral imagery and ensemble learning [J/OL]. Remote. Sens., 2020, 12(12): 2028 [2023-11-15]. .
44	GE H X, MA F, LI Z W, et al.. Improved accuracy of phenological detection in rice breeding by using ensemble models of machine learning based on UAV-RGB imagery [J/OL]. Remote. Sens., 2021, 13(14):2678 [2023-11-15]. .
45	PRAGATHI K. Crop yield prediction and fertilizer recommendation using voting based ensemble classifier [J]. Int. J. Innovat. Res. Technol., 2021, 8(6): 510-516.

[1]	Yanfei FANG, Xiaoying LUO, Jianghua TANG, Tingting SUN, Luzhen WANG, Tian TANG, Wenxiu XU. Effects of Sowing Methods on Yield， Dry Matter and Water Use Efficiency of Spring Wheat in Dryland [J]. Journal of Agricultural Science and Technology, 2024, 26(1): 173-181.
[2]	Rui TIAN, Hua ZHANG, Meihong HUANG, Zhenqi SHAO, Xihuan LI, Caiying ZHANG. Mining of Candidate Genes and Genetic Loci Conferring Drought Tolerance in Soybean [J]. Journal of Agricultural Science and Technology, 2023, 25(9): 69-82.
[3]	JIAN Tiancai, KANG Jianhong, WU Hongliang, LIU Genhong, GAO Di, MA Xueying, LI Xin. Antioxidative Characteristics Study of Nitrogen in Alleviating Premature Senescence of Spring Wheat at High Temperature after Anthesis [J]. Journal of Agricultural Science and Technology, 2021, 23(7): 33-44.
[4]	YANG Jing, LIAO Guiping*, LIU Fan, GUAN Chunyun. Prediction of Chlorophyll Content of Rape Leaves with Hyperspectral Imaging Technology [J]. Journal of Agricultural Science and Technology, 2020, 22(5): 86-96.
[5]	WANG Yi, YAN Zhiyong*. Detection System of Chlorophyll Content of Cyclobalanopsis glauca Using Image Processing Technology [J]. Journal of Agricultural Science and Technology, 2017, 19(4): 59-64.
[6]	BA0 Shun-shu, HE Dong-xian, GU0 Shun-xing. Suitable Photoperiod of Dendrobium Officinale in \|vitro under Controlled Environment [J]. , 2007, 9(6): 90-94.
[7]	ZHOU Yi, WANG Chuan-jiang, CAO Yi-ping . Drought Resistance Effect and Analysis of Spraying Salt Nitro-fulvic Acid on Spring Wheat [J]. , 2005, 7(4): 46-50.