基于高光谱遥感和集成学习方法的冬小麦产量估测研究

doi:10.3864/j.issn.0578-1752.2021.16.005

摘要/Abstract

摘要：

【目的】利用2种灌溉处理下不同发育阶段的冬小麦冠层高光谱信息,通过机器学习方法对小麦籽粒产量进行估测精度研究,明确产量最佳估测模型,对于育种工作有着重要应用价值。【方法】以黄淮麦区207个主栽小麦品种为材料,于2018—2019和2019—2020年度连续2个生长季在河南省新乡基地的正常灌溉和节水处理下种植,并调查开花期、灌浆前期和灌浆中期的冠层高光谱数据,分别以6种机器学习方法和集成方法建立光谱指数产量估测模型。【结果】2种灌溉处理下,3个生育期各光谱指数均与产量呈极显著相关（P<0.0001）,且表现出较高的遗传力（0.61-0.85）,主要受遗传因素控制。在正常灌溉处理下,与传统机器学习方法表现最佳的模型相比,集成学习方法在3个生育期的平均决定系数(R²) 分别由0.610、0.611和0.640提高至0.649、0.612和0.675,平均均方根误差 (RMSE) 分别降低至0.607、0.612和0.593 t·hm^-2;节水处理下,3个生育期的平均R²分别由0.461、0.408和0.452提高至0.467、0.433和0.498,平均RMSE分别降低至0.519、0.559和0.504 t·hm^-2。【结论】利用集成方法将不同模型估测结果进行结合,能够有效地提高产量估测精度,2种灌溉处理下均在灌浆中期估测精度最佳,可为冬小麦育种工作中产量估测提供参考。

关键词: 冬小麦, 产量, 高光谱, 集成方法, 机器学习

Abstract:

【Objective】Using the hyperspectral data of winter wheat canopy at different development stages under two irrigation treatments, the estimation accuracy of wheat grain yield was studied by machine learning method, and the best yield estimation model was defined, which had the important application value for crop breeding. 【Method】 A total of 207 widely-grown wheat varieties in the Yellow and Huai Valleys Winter Wheat Zone (YHVWWZ) of China were planted under full irrigation and limited irrigation treatments in Xinxiang, Henan province during two consecutive growing seasons of 2018-2019 and 2019-2020, the canopy hyperspectral was investigated at three growth stages after flowering, and six machine learning methods and ensemble methods were adopted to establish yield prediction model by using spectral index as input features.【Result】 The spectral indices at each growth stage were significantly correlated with yield (P<0.0001) under both the two irrigation treatments, and also showed high heritability (0.61-0.85) across all the three growth stages under both the irrigation treatment, which were mainly controlled by genetic factors. Under the full irrigation treatment, compared with the model with the best performance of traditional machine learning methods, the average coefficient of determination (R²) of ensemble learning method in the three growth stages increased from 0.610, 0.611 and 0.640 to 0.649, 0.612 and 0.675, respectively, and the average root mean square error (RMSE) decreased to 0.607, 0.612 and 0.593 t·hm^-2, respectively; Under the limited irrigation treatment, the average R ² increased from 0.461, 0.408 and 0.452 to 0.467, 0.433 and 0.498, respectively, and the average RMSE decreased to 0.519, 0.559 and 0.504 t·hm ^-2, respectively.【Conclusion】Combining the prediction results of different models with the ensemble learning method could effectively improve the yield estimation accuracy, and the mid grain filling achieved the best prediction accuracy under both the two irrigation treatments. Overall, this study could provide the reference for yield estimation in winter wheat breeding.

Key words: winter wheat, grain yield, hyperspectral, ensemble method, machine learning

费帅鹏,禹小龙,兰铭,李雷,夏先春,何中虎,肖永贵. 基于高光谱遥感和集成学习方法的冬小麦产量估测研究[J]. 中国农业科学, 2021, 54(16): 3417-3427.

FEI ShuaiPeng,YU XiaoLong,LAN Ming,LI Lei,XIA XianChun,HE ZhongHu,XIAO YongGui. Research on Winter Wheat Yield Estimation Based on Hyperspectral Remote Sensing and Ensemble Learning Method[J]. Scientia Agricultura Sinica, 2021, 54(16): 3417-3427.

0
/ / 推荐

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: https://www.chinaagrisci.com/CN/10.3864/j.issn.0578-1752.2021.16.005

https://www.chinaagrisci.com/CN/Y2021/V54/I16/3417

图/表 9

表1

本研究选用的光谱指数"

光谱指数 Spectral index	名称 Name	公式 Formula
NDVI^[23]	归一化光谱指数 Normalized difference vegetation index	$\frac{R_{800}-R_{670}}{R_{800}+R_{670}}$
MCARI^[24]	修正叶绿素吸收比指数 Modified chlorophyll absorption ratio index	$\left[\left(R_{702}-R_{671}\right)-0.2\left(R_{702}-R_{549}\right)\right] \times \frac{R_{702}}{R_{671}}$
NDRE^[25]	归一化红边光谱指数 Normalized difference red edge	$\frac{R_{790}-R_{720}}{R_{790}+R_{720}}$
GNDVI^[26]	绿色归一化光谱指数 Green normalized difference vegetation index	$\frac{R_{750}-R_{550}}{R_{750}+R_{550}}$
MSR^[23]	修正红边比值指数 Modified simple ratio index	$\frac{R_{750} / R_{705}-1}{\sqrt{R_{750} / R_{705}+1}}$
NDRSR^[27]	归一化红边简单比值指数 Normalized difference red-edge simple ratio	$\frac{R_{872}-R_{712}}{R_{872}+R_{712}}$
MTVI^[28]	修正三角光谱指数 Modified triangular vegetation index	1.2[1.2(R₈₀₀-R₅₀₀)-2.6(R₆₇₀-R₅₅₀)]
MTCI2^[29]	MERIS陆地叶绿素指数2 MERIS terrestrial chlorophyll index 2	$\frac{R_{754}-R_{709}}{R_{709}+R_{681}}$
MNDVI^[30]	修正归一化光谱指数 Modified normalized difference vegetation index	$\frac{R_{750}-R_{705}}{R_{750}+R_{705}-2 R_{445}}$
RDVI^[31]	重归一化光谱指数 Renormalized difference vegetation index	$\frac{R_{800}-R_{670}}{\sqrt{R_{800}+R_{670}}}$
VDI^[32]	植被干指数 Vegetation dry index	$\frac{R_{970}-R_{900}}{R_{970}+R_{900}}$
CI^[33]	叶绿素指数 Chlorophyll index	(R₇₄₉-R₇₂₀)-(R₇₀₁-R₆₇₂)
VREI^[34]	沃格尔曼红边指数 Vogelmann red edge index	$\frac{R_{742}}{R_{722}}$
ARVI^[35]	大气抗性光谱指数 Atmospherically resistant vegetation index	$\frac{R_{872}-\left[R_{661}-\left(R_{488}-R_{661}\right)\right]}{R_{872}+\left[R_{661}-\left(R_{488}-R_{661}\right)\right]}$
NDMI^[36]	归一化物质指数 Normalized difference matter index	$\frac{R_{1649}-R_{1792}}{R_{1649}+R_{1792}}$

表1

图1

表2

表3

图2

图3

图4

图5

表4

参考文献 44

[1]	HERNANDEZ J, LOBOS G A, MATUS I, DEL POZO A, SILVA P, GALLEGUILLOS M. Using ridge regression models to estimate grain yield from field spectral data in bread wheat (Triticum aestivum L. ) grown under three water regimes. Remote Sensing, 2015, 7(2):2109-2126. doi: 10.3390/rs70202109
[2]	MONTESINOS-LÓPEZ O A, MONTESINOS-LÓPEZ A, CROSSA J, DELOS G, CAMPOS , ALVARADO G, SUCHISMITA M, RUTKOSKI J, GONZÁLEZ-PÉREZ L, BURGUEÑO J. Predicting grain yield using canopy hyperspectral reflectance in wheat breeding data. Plant Methods, 2017, 13:4. doi: 10.1186/s13007-016-0154-2
[3]	HASSAN M A, YANG M, RASHEED A, YANG G, REYNOLDS M, XIA X, XIAO Y, HE Z. A rapid monitoring of NDVI across the wheat growth cycle for grain yield prediction using a multi-spectral UAV platform. Plant Science, 2019, 282:95-103. doi: 10.1016/j.plantsci.2018.10.022
[4]	HASSAN M A, YANG M, RASHEED A, JIN X, XIA X, XIAO Y, HE Z. Time-series multispectral indices from unmanned aerial vehicle imagery reveal senescence rate in bread wheat. Remote Sensing, 2018, 10(6):809. doi: 10.3390/rs10060809
[5]	GITELSON A A, PENG Y, ARKEBAUER T J, SCHEPERS J. Relationships between gross primary production, green LAI, and canopy chlorophyll content in maize: Implications for remote sensing of primary production. Remote Sensing of Environment, 2014, 144:65-72. doi: 10.1016/j.rse.2014.01.004
[6]	BOLTON D K, FRIEDL M A. Forecasting crop yield using remotely sensed vegetation indices and crop phenology metrics. Agricultural & Forest Meteorology, 2013, 173:74-84.
[7]	李岚涛, 李静, 明金, 汪善勤, 任涛, 鲁剑巍. 冬油菜叶面积指数高光谱监测最佳波宽与有效波段研究. 农业机械学报, 2018, 49(2):156-165.
	LI L T, LI J, MING J, WANG S Q, REN T, LU J W. Selection optimization of hyperspectral bandwidth and effective wavelength for predicting leaf areaindex in winter oilseed rape. Transactions of the Chinese Society for Agricultural Machinery, 2018, 49(2):156-165. (in Chinese)
[8]	SHAH S H, ANGEL Y, HOUBORG R, ALI S, MCCABE M F. A random forest machine learning approach for the retrieval of leaf chlorophyll content in wheat. Remote Sensing, 2019, 11:920. doi: 10.3390/rs11080920
[9]	BREIMANL. Random forests. Machine Learning, 2001, 45:5-32. doi: 10.1023/A:1010933404324
[10]	SAIN, STEPHAN R. The nature of statistical learning theory. Technometrics, 1996, 38:409.
[11]	BRADLEY J B. Neural networks: A comprehensive foundation. Information Processing & Management, 1995, 31:786. doi: 10.1016/0306-4573(95)90003-9
[12]	WANG L, ZHOU X, ZHU X, DONG Z, GUO W. Estimation of biomass in wheat using random forest regression algorithm and remote sensing data. The Crop Journal, 2016, 4:212-219. doi: 10.1016/j.cj.2016.01.008
[13]	YUAN H, YANG G, LI C, WANG Y, LIU J, YU H, FENG H, XU B, ZHAO X, YANG X. Retrieving soybean leaf area index from unmanned aerial vehicle hyperspectral remote sensing: Analysis of RF, ANN, and SVM regression models. Remote Sensing, 2017, 9:309. doi: 10.3390/rs9040309
[14]	JIN X, XU X, SONG X, LI Z, WANG J, GUO W. Estimation of leaf water content in winter wheat using grey relational analysis-partial least squares modeling with hyperspectral data. Agronomy Journal, 2013, 105:1385-1392. doi: 10.2134/agronj2013.0088
[15]	FENG L, ZHANG Z, MA Y, DU Q, WILLIAMS P, DREWRY J, LUCK B. Alfalfa yield prediction using UAV-based hyperspectral imagery and ensemble learning. Remote Sensing, 2020, 12(12):2028. doi: 10.3390/rs12122028
[16]	WOLPERT D H. Stacked generalization. Neural Networks, 1992, 5:241-259. doi: 10.1016/S0893-6080(05)80023-1
[17]	TING K M, WITTEN I H. Issues in stacked generalization. Journal of Artificial Intelligence Research, 1999, 10:271-289. doi: 10.1613/jair.594
[18]	FU P, MEACHAM-HENSOLD K, GUAN K, BERNACCHI C J. Hyperspectral leaf reflectance as proxy for photosynthetic capacities: An ensemble approach based on multiple machine learning algorithms. Frontiers in Plant Science, 2019, 10.
[19]	HEALEY S P, COHEN W B, YANG Z, KENNETH BREWER C, BROOKS E B, GORELICK N, HERNANDEZ A J, HUANG C, JOSEPH HUGHES M, KENNEDY R E, LOVELAND T R, MOISEN G G, SCHROEDER T A, STEHMAN S V, VOGELMANN J E, WOODCOCK C E, YANG L, ZHU Z. Mapping forest change using stacked generalization: An ensemble approach. Remote Sensing of Environment, 2018, 204:717-728. doi: 10.1016/j.rse.2017.09.029
[20]	WILLIAMSCK, RASMUSSENCE. Gaussian processes for machine learning. Cambridge, CA: MIT Press, 2006.
[21]	MCDONALD G C. Ridge regression. Wiley Interdisciplinary Reviews Computational Statistics, 2009, 1:93-100. doi: 10.1002/wics.v1:1
[22]	LIANG L, DI L P, ZHANG L P, DENG M X, QIN Z H, ZHAO S H, LIN H. Estimation of crop LAI using hyperspectral vegetation indices and a hybrid inversion method. Remote Sensing of Environment, 2015, 165:123-134. doi: 10.1016/j.rse.2015.04.032
[23]	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. Remote Sensing of Environment, 2002, 81(2/3):337-354. doi: 10.1016/S0034-4257(02)00010-X
[24]	DAUGHTRY C S T, WALTHALL C L, KIM M S, DE COLSTOUN E B, MCMURTREY J E. Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance. Remote Sensing of Environment, 2000, 74:229-239. doi: 10.1016/S0034-4257(00)00113-9
[25]	RODRIGUEZ D, FITZGERALD G J, BELFORD R, CHRISTENSEN L K. Detection of nitrogen deficiency in wheat from spectral reflectance indices and basic crop eco-physiological concepts. Australian Journal of Agricultural Research, 2006, 57:781-789. doi: 10.1071/AR05361
[26]	GITELSON A A, KAUFMAN Y J, MERZLYAK M N. Use of a green channel in remote sensing of global vegetation from EOS-MODIS. Remote Sensing of Environment, 1996, 58:289-298. doi: 10.1016/S0034-4257(96)00072-7
[27]	GITELSON A A, VINA A, CIGANDA V, RUNDQUIST D C, ARKEBAUER T J. Remote estimation of canopy chlorophyll content in crops. Geophysical Research Letters, 2005, 32:1-4.
[28]	HABOUDANE D, MILLER J R, PATTEY E, ZARCO-TEJADA P J, STRACHAN I B. Hyperspectral vegetation indices and novel algorithms for predicting green LAI of crop canopies: Modeling and validation in the context of precision agriculture. Remote Sensing of Environment, 2004, 90:337-352. doi: 10.1016/j.rse.2003.12.013
[29]	DASH J, CURRAN P J. Evaluation of the meris terrestrial chlorophyll index (MTCI). Advances in Space Research, 2007, 39:100-104. doi: 10.1016/j.asr.2006.02.034
[30]	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. Remote Sensing of Environment, 2002, 81:337-354. doi: 10.1016/S0034-4257(02)00010-X
[31]	ROUJEAN J L, BREON F M. Estimating PAR absorbed by vegetation from bidirectional reflectance measurements. Remote Sensing of Environment, 1995, 51:375-384. doi: 10.1016/0034-4257(94)00114-3
[32]	PENUELAS J, FILELLA I, BIEL C S, SERRANO L, SAVE R. The Reflectance at the 950-970 Nm region as an indicator of plant water status. International Journal of Remote Sensing, 1993, 14(10):1887-1905. doi: 10.1080/01431169308954010
[33]	GUPTA R K, VIJAYAN D, PRASAD T S. New hyperspectral vegetation characterization parameters. Advances in Space Research, 2001, 28(1):201-206. doi: 10.1016/S0273-1177(01)00346-5
[34]	VOGELMANN J, ROCK B, MOSS D. Red edge spectral measurements from sugar maple leaves. Remote Sensing. 1993, 14:1563-1575.
[35]	KAUFMAN Y J, TANRE D. Atmospherically resistant vegetation index (ARVI) for eos-modis. IEEE Transactions on Geoscience and Remote Sensing, 1992, 30:261-270. doi: 10.1109/36.134076
[36]	WANG L, HUNT E R, JR, QU J J, HAO X, DAUGHTRY C S T. Towards estimation of canopy foliar biomass with spectral reflectance measurements. Remote Sensing of Environment, 2011, 115(3):836-840. doi: 10.1016/j.rse.2010.11.011
[37]	周志华. 机器学习.第一版. 北京: 清华大学出版社, 2016: 181-182.
	ZHOU Z H. Machine Learning.1st edition. Beijing: Tsinghua University Press, 2016: 181-182. (in Chinese)
[38]	邓威, 郭钇秀, 李勇, 朱亮, 刘定国. 基于特征选择和Stacking集成学习的配电网网损估测. 电力系统保护与控制, 2020, 48:108-115.
	DENG W, GUO Y X, LI Y, ZHU L, LIU D G. Power losses prediction based on feature selection and Stacking integrated learning. Power System Protection and Control, 2020, 48:108-115. (in Chinese)
[39]	JULIANE B, ANDREAS B, SIMON B, JANIS B, SILAS E, GEORG B. Estimating biomass of barley using crop surface models (CSMs) derived from UAV-Based RGB imaging. Remote Sensing, 2014, 6(11):10395-10412. doi: 10.3390/rs61110395
[40]	ZOU X C, MOTTUS M. Sensitivity of common vegetation indices to the canopy structure of field crops. Remote Sensing, 2017, 9:994. doi: 10.3390/rs9100994
[41]	FENG L, LI Y, WANG Y, DU Q. Estimating hourly and continuous ground-level PM2.5 concentrations using an ensemble learning algorithm: The ST-Stacking model. Atmospheric Environment, 2020, 223:117242. doi: 10.1016/j.atmosenv.2019.117242
[42]	FRAME J, MERRILEES D W. The effect of tractor wheel passes on herbage production from diploid and tetraploid ryegrass swards. Grass and Forage Science, 1996, 51:13-20. doi: 10.1111/gfs.1996.51.issue-1
[43]	VAN D L, M J, POLLEY E C, HUBBARDAE. Super learner. Statistical Applications in Genetics & Molecular Biology, 2007, 6(1):25.
[44]	陈智芳, 宋妮, 王景雷, 孙景生. 基于高光谱遥感的冬小麦叶水势估测模型. 中国农业科学, 2017, 50(5):871-880.
	CHEN Z F, SONG N, WANG J L, SUN J S. Leaf water potential estimating models of winter wheat based on hyperspectral remote sensing. Scientia Agricultura Sinica, 2017, 50(5):871-880. (in Chinese)

光谱指数 Spectral index	开花期Flowering		灌浆前期Early grain filling		灌浆中期Mid grain filling
光谱指数 Spectral index	\|r\|	H²	\|r\|	H²	\|r\|	H²
NDVI	0.50***	0.83	0.53***	0.75	0.66***	0.74
MCARI	0.61***	0.85	0.65***	0.85	0.69***	0.82
NDRE	0.71***	0.81	0.65***	0.79	0.72***	0.78
GNDVI	0.68***	0.82	0.63***	0.78	0.71***	0.77
MSR	0.65***	0.80	0.62***	0.76	0.70***	0.76
NDRSR	0.72***	0.82	0.67***	0.79	0.73***	0.79
MTVI	0.59***	0.77	0.60***	0.73	0.63***	0.75
MTCI2	0.63***	0.83	0.59***	0.80	0.68***	0.80
MNDVI	0.62***	0.83	0.63***	0.76	0.69***	0.83
RDVI	0.60***	0.77	0.62***	0.77	0.66***	0.78
VDI	0.45***	0.84	0.43***	0.80	0.62***	0.83
CI	0.61***	0.82	0.59***	0.79	0.64***	0.81
VREI	0.69***	0.81	0.65***	0.75	0.72***	0.76
ARVI	0.52***	0.79	0.54***	0.73	0.65***	0.73
NDMI	0.53***	0.80	0.56***	0.63	0.61***	0.77

光谱指数 Spectral index	开花期 Flowering		灌浆前期 Early grain filling		灌浆中期 Mid grain filling
光谱指数 Spectral index	\|r\|	H²	\|r\|	H²	\|r\|	H²
NDVI	0.48***	0.63	0.42***	0.69	0.53***	0.65
MCARI	0.57***	0.64	0.49***	0.70	0.56***	0.68
NDRE	0.56***	0.67	0.48***	0.78	0.55***	0.73
GNDVI	0.54***	0.66	0.41***	0.73	0.54***	0.71
MSR	0.55***	0.65	0.43***	0.74	0.52***	0.69
NDRSR	0.57***	0.68	0.49***	0.79	0.55***	0.73
MTVI	0.45***	0.64	0.46***	0.68	0.49***	0.68
MTCI2	0.50***	0.62	0.45***	0.71	0.50***	0.66
MNDVI	0.52***	0.64	0.49***	0.67	0.51***	0.71
RDVI	0.49***	0.66	0.47***	0.65	0.50***	0.67
VDI	0.59***	0.69	0.61***	0.68	0.58***	0.73
CI	0.48***	0.64	0.49***	0.74	0.52***	0.74
VREI	0.55***	0.68	0.48***	0.68	0.52***	0.69
ARVI	0.49***	0.61	0.42***	0.71	0.51***	0.68
NDMI	0.44***	0.73	0.49***	0.66	0.50***	0.72

模型 Model	正常灌溉处理 Full irrigation treatment			节水处理 Limited irrigation treatment
模型 Model	开花期 Flowering	灌浆前期 Early grain filling	灌浆中期 Mid grain filling	开花期 Flowering	灌浆前期 Early grain filling	灌浆中期 Mid grain filling
ANN	-4.53	0.43	-0.59	-4.10	-2.81	0.93
GP	0.24	0.26	-1.85	2.10	0.77	0.86
MLR	1.91	-0.19	4.51	-2.14	-0.14	-2.02
RF	0.84	-0.30	-7.30	-0.47	0.02	-0.37
RR	6.13	0.61	4.32	3.20	4.06	3.28
SVM	-3.41	0.42	2.34	2.43	-0.56	-1.59