基于随机森林回归的油菜叶片SPAD值遥感估算
|
摘要
以西北地区典型经济作物油菜为研究对象,利用SVC-1024i型便携式光谱仪和SPAD-502型叶绿素仪测定了油菜不同生育期的叶片光谱反射率和SPAD值。通过分析油菜原始光谱及10种光谱指数与SPAD值的相关关系,基于光谱指数构建了不同生育期油菜叶片SPAD值随机森林回归(RF)估算模型,并利用独立样本对所建模型进行验证,同时结合传统的一元线性回归模型和多元逐步回归模型与其进行比较。结果表明:油菜叶片SPAD值在全生育期内呈现出先上升后下降的趋势;各光谱指数在不同生育期及全生育期与SPAD值的相关性均达到0.01水平的显著相关;基于光谱指数构建的随机森林回归模型在油菜各个生育期及全生育期建模和预测结果明显优于同期的传统回归模型,建模R2达0.90以上,验证R2达0.81以上,RMSE在1.571～5.004,RE在2.66%～13.22%,是油菜叶片SPAD值的最优估算模型。
The chlorophyll content is an importance parameter for evaluating crop growth and real-time non-destructive and quick estimation of leaf chlorophyll content can provide important information about plant stress,nutritional status,and relationships between plants and their environment,which has great significance in guiding agricultural production and improving crop yield. The conclusions are as follows: 1) With the increase of SPAD value,the leaf spectral reflectance decreased in the visible light region,and the "red shift"phenomena were detected at the red edge position. 2) The correlations between rapeseed leaf SPAD values and spectral indices were significant in all growth periods. Among the spectral indices,TCARI,GRVI,and NPCI were negatively correlated with leaf SPAD values,while the rest of spectral indices were positively correlated with the leaf SPAD values. 3) Leaf SPAD value estimation models using the traditional simple linear regression method,multiple stepwise regression method,and Random Forest method base on the spectral indices all passed the significance tests. In order to evaluate each model's estimation accuracy and to further compare the performances of the three models for each stage,the coefficient of determination( R2) of each model was calculated respectively for both modeling sets and validation sets.The results indicated that random forest model had the best modeling and verification accuracy in each growth period with the coefficient of determination higher than 0.91 for the modeling and greater than 0.74 for the validation set. Therefore,Random Forest Model is an optimal model for estimating rapeseed leaf SPAD values and may provide a theoretical basis and technical support to improve remote sensing inversion accuracy of rapeseed chlorophyll content.
The chlorophyll content is an importance parameter for evaluating crop growth and real-time non-destructive and quick estimation of leaf chlorophyll content can provide important information about plant stress,nutritional status,and relationships between plants and their environment,which has great significance in guiding agricultural production and improving crop yield. The conclusions are as follows: 1) With the increase of SPAD value,the leaf spectral reflectance decreased in the visible light region,and the "red shift"phenomena were detected at the red edge position. 2) The correlations between rapeseed leaf SPAD values and spectral indices were significant in all growth periods. Among the spectral indices,TCARI,GRVI,and NPCI were negatively correlated with leaf SPAD values,while the rest of spectral indices were positively correlated with the leaf SPAD values. 3) Leaf SPAD value estimation models using the traditional simple linear regression method,multiple stepwise regression method,and Random Forest method base on the spectral indices all passed the significance tests. In order to evaluate each model's estimation accuracy and to further compare the performances of the three models for each stage,the coefficient of determination( R2) of each model was calculated respectively for both modeling sets and validation sets.The results indicated that random forest model had the best modeling and verification accuracy in each growth period with the coefficient of determination higher than 0.91 for the modeling and greater than 0.74 for the validation set. Therefore,Random Forest Model is an optimal model for estimating rapeseed leaf SPAD values and may provide a theoretical basis and technical support to improve remote sensing inversion accuracy of rapeseed chlorophyll content.
引文
[1]丁希斌,刘飞,张初,等.基于高光谱成像技术的油菜叶片SPAD值检测[J].光谱学与光谱分析,2015,35(2):486-491.
[2]徐新刚,赵春江,王纪华,等.新型光谱曲线特征参数与水稻叶绿素含量间的关系研究[J].光谱学与光谱分析,2011,31(1):188-191.
[3]王磊,白由路.不同氮处理春玉米叶片光谱反射率与叶片全氮和叶绿素含量的相关研究[J].中国农业科学,2005,38(11):2268-2276.
[4]Dash J,Curran P J.The MERIS terrestrial chlorophyll index[J].International journal of remote sensing,2004,25(23):5403-5413.
[5]Gitelson A A,Vina A,Ciganda V,et al.Remote estimation of canopy chlorophyll content in crops[J].Geophysical Research Letters,2005,32(8):93-114.
[6]Broge N H.,Mortensen J V.Deriving green crop area index and canopy chlorophyll density of winter wheat from spectral reflectance data[J].Remote Sensing of Environment,2002,81(1):45-57.
[7]姚付启,张振华,杨润亚,等.基于主成分分析和BP神经网络的法国梧桐叶绿素含量高光谱反演研究[J].测绘科学,2010,35(1):109-112.
[8]宫兆宁,赵雅莉,赵文吉,等.基于光谱指数的植物叶片叶绿素含量的估算模型[J].生态学报,2014,34(20):5736-5745.
[9]Grimm R,Behrens T,Maerker M,et al.Soil organic carbon concentrations and stocks on Barro Colorado Island-Digital soil mapping using Random Forests analysis[J].Geoderma,2008,146(1-2):102-113.
[10]Mutanga O,Adam E,Azong Cho M.High density biomass estimation for wetland vegetation using World View-2 imagery and random forest regression algorithm[J].International Journal of Applied Earth Observation and Geoinformation,2012,18(1):399-406.
[11]Colin J,Jungho Im.Forest biomass estimation from airborne LADARdata using machine learning approaches[J].Remote Sensing of Environment,2012,125(125):80-91.
[12]李粉玲,王力,刘京,等.基于高分一号卫星数据的冬小麦叶片SPAD值遥感估算[J].农业机械学报,2015,46(9):273-281.
[13]王丽爱,马昌,周旭东,等.基于随机森林回归算法的小麦叶片SPAD值遥感估算[J].农业机械学报,2015,46(1):259-265.
[14]韩兆迎,朱西存,房贤一,等.基于SVM与RF的苹果树冠LAI高光谱估测[J].光谱学与光谱分析,2016,36(3):800-805.
[15]赵佳佳,冯美臣,王超,等.基于光谱植被指数的冬小麦叶绿素含量反演[J].山西农业大学学报(自然科学版),2014,34(5):391-396.
[16]Pearson R L,Miller L D.Remote Mapping of Standing Crop Biomass for Estimation of the Productivity of the Shortgrass Prairie[C]//Proceedings of the Eighth International Symposium on Remote Sensing of Environment.Michigan USA:Environmental Research Institute of Michigan,1972:7-12.
[17]Gitelson A,Merzlyak M N.Spectral reflectance changes associated with autumn senescence of aesculus hippocastanum L.and acer platanoides L.leaves-Spectral features and relation to chlorophyll estimation[J].Journal of Plant Physiology,1994,143(3):286-292.
[18]Datt B.A new reflectance index for remote sensing of chlorophyll content in higher plants:Tests using Eucalyptus leaves[J].Journal of Plant Physiology,1999,154(1):30-36.
[19]Haboudane D,Miller J R.,Tremblay N,et al.Integrated narrowband vegetation indices for prediction of crop chlorophyll content for application to precision agriculture[J].Remote Sensing of Environment,2002,81(2-3):416-426.
[20]Gitelson A A,Kaufman Y J,Merzlyak M N.Use of a green channel in remote sensing of global vegetation from EOS-MODIS[J].Remote Sensing of Environment,1996,58(3):289-298.
[21]Blackburn G A.Spectral indices for estimating photosynthetic pigment concentrations:a test using senescent tree leaves[J].International Journal of Remote Sensing,1998,19(4):657-675.
[22]Chen Pengfei,Haboudane D,Tremblay N,et al.New spectral indicator assessing the efficiency of crop nitrogen treatment in corn and wheat[J].Remote Sensing of Environment,2010,114(9):1987-1997.
[23]梁亮,杨敏华,张连蓬,等.基于SVR算法的小麦冠层叶绿素含量高光谱反演[J].农业工程学报,2012,28(20):162-171.
[24]孙雪莲,舒清态,欧光龙,等.基于随机森林回归模型的思茅松人工林生物量遥感估测[J].林业资源管理,2015,(01):71-76.
[25]方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛,2011,26(3):32-38.
[26]王丽爱,周旭东,朱新开,等.基于HJ-CCD数据和随机森林算法的小麦叶面积指数反演[J].农业工程学报,2016,32(3):149-154.
[27]张文英,王凯华.甘蓝型油菜SPAD值与叶绿素含量关系分析[J].中国农学通报,2012,28(21):92-95.
[28]王瑞,陈永忠,陈隆升,等.油茶叶片SPAD值与叶绿素含量的相关分析[J].中南林业科技大学学报,2013,33(2):77-80.
[29]艾天成,李方敏,周治安,等.作物叶片叶绿素含量与SPAD值相关性研究[J].湖北农学院学报,2000,20(1):6-8.
[31]罗丹,常庆瑞,齐雁冰,等.基于光谱指数的冬小麦冠层叶绿素含量估算模型研究[J].麦类作物学报,2016,36(9):1225-1233.
[32]潘蓓,赵庚星,朱西存,等.利用高光谱植被指数估测苹果树冠层叶绿素含量[J].光谱学与光谱分析,2013,33(8):2203-2206.
[33]梁栋,管青松,黄文江,等.基于支持向量机回归的冬小麦叶面积指数遥感反演[J].农业工程学报,2013,29(7):117-123.
[34]Cutler D R,Edwards T C,Beard K H,et al.Random forests for classification in ecology[J].Ecology,2007,88(11):2783.
[2]徐新刚,赵春江,王纪华,等.新型光谱曲线特征参数与水稻叶绿素含量间的关系研究[J].光谱学与光谱分析,2011,31(1):188-191.
[3]王磊,白由路.不同氮处理春玉米叶片光谱反射率与叶片全氮和叶绿素含量的相关研究[J].中国农业科学,2005,38(11):2268-2276.
[4]Dash J,Curran P J.The MERIS terrestrial chlorophyll index[J].International journal of remote sensing,2004,25(23):5403-5413.
[5]Gitelson A A,Vina A,Ciganda V,et al.Remote estimation of canopy chlorophyll content in crops[J].Geophysical Research Letters,2005,32(8):93-114.
[6]Broge N H.,Mortensen J V.Deriving green crop area index and canopy chlorophyll density of winter wheat from spectral reflectance data[J].Remote Sensing of Environment,2002,81(1):45-57.
[7]姚付启,张振华,杨润亚,等.基于主成分分析和BP神经网络的法国梧桐叶绿素含量高光谱反演研究[J].测绘科学,2010,35(1):109-112.
[8]宫兆宁,赵雅莉,赵文吉,等.基于光谱指数的植物叶片叶绿素含量的估算模型[J].生态学报,2014,34(20):5736-5745.
[9]Grimm R,Behrens T,Maerker M,et al.Soil organic carbon concentrations and stocks on Barro Colorado Island-Digital soil mapping using Random Forests analysis[J].Geoderma,2008,146(1-2):102-113.
[10]Mutanga O,Adam E,Azong Cho M.High density biomass estimation for wetland vegetation using World View-2 imagery and random forest regression algorithm[J].International Journal of Applied Earth Observation and Geoinformation,2012,18(1):399-406.
[11]Colin J,Jungho Im.Forest biomass estimation from airborne LADARdata using machine learning approaches[J].Remote Sensing of Environment,2012,125(125):80-91.
[12]李粉玲,王力,刘京,等.基于高分一号卫星数据的冬小麦叶片SPAD值遥感估算[J].农业机械学报,2015,46(9):273-281.
[13]王丽爱,马昌,周旭东,等.基于随机森林回归算法的小麦叶片SPAD值遥感估算[J].农业机械学报,2015,46(1):259-265.
[14]韩兆迎,朱西存,房贤一,等.基于SVM与RF的苹果树冠LAI高光谱估测[J].光谱学与光谱分析,2016,36(3):800-805.
[15]赵佳佳,冯美臣,王超,等.基于光谱植被指数的冬小麦叶绿素含量反演[J].山西农业大学学报(自然科学版),2014,34(5):391-396.
[16]Pearson R L,Miller L D.Remote Mapping of Standing Crop Biomass for Estimation of the Productivity of the Shortgrass Prairie[C]//Proceedings of the Eighth International Symposium on Remote Sensing of Environment.Michigan USA:Environmental Research Institute of Michigan,1972:7-12.
[17]Gitelson A,Merzlyak M N.Spectral reflectance changes associated with autumn senescence of aesculus hippocastanum L.and acer platanoides L.leaves-Spectral features and relation to chlorophyll estimation[J].Journal of Plant Physiology,1994,143(3):286-292.
[18]Datt B.A new reflectance index for remote sensing of chlorophyll content in higher plants:Tests using Eucalyptus leaves[J].Journal of Plant Physiology,1999,154(1):30-36.
[19]Haboudane D,Miller J R.,Tremblay N,et al.Integrated narrowband vegetation indices for prediction of crop chlorophyll content for application to precision agriculture[J].Remote Sensing of Environment,2002,81(2-3):416-426.
[20]Gitelson A A,Kaufman Y J,Merzlyak M N.Use of a green channel in remote sensing of global vegetation from EOS-MODIS[J].Remote Sensing of Environment,1996,58(3):289-298.
[21]Blackburn G A.Spectral indices for estimating photosynthetic pigment concentrations:a test using senescent tree leaves[J].International Journal of Remote Sensing,1998,19(4):657-675.
[22]Chen Pengfei,Haboudane D,Tremblay N,et al.New spectral indicator assessing the efficiency of crop nitrogen treatment in corn and wheat[J].Remote Sensing of Environment,2010,114(9):1987-1997.
[23]梁亮,杨敏华,张连蓬,等.基于SVR算法的小麦冠层叶绿素含量高光谱反演[J].农业工程学报,2012,28(20):162-171.
[24]孙雪莲,舒清态,欧光龙,等.基于随机森林回归模型的思茅松人工林生物量遥感估测[J].林业资源管理,2015,(01):71-76.
[25]方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛,2011,26(3):32-38.
[26]王丽爱,周旭东,朱新开,等.基于HJ-CCD数据和随机森林算法的小麦叶面积指数反演[J].农业工程学报,2016,32(3):149-154.
[27]张文英,王凯华.甘蓝型油菜SPAD值与叶绿素含量关系分析[J].中国农学通报,2012,28(21):92-95.
[28]王瑞,陈永忠,陈隆升,等.油茶叶片SPAD值与叶绿素含量的相关分析[J].中南林业科技大学学报,2013,33(2):77-80.
[29]艾天成,李方敏,周治安,等.作物叶片叶绿素含量与SPAD值相关性研究[J].湖北农学院学报,2000,20(1):6-8.
[31]罗丹,常庆瑞,齐雁冰,等.基于光谱指数的冬小麦冠层叶绿素含量估算模型研究[J].麦类作物学报,2016,36(9):1225-1233.
[32]潘蓓,赵庚星,朱西存,等.利用高光谱植被指数估测苹果树冠层叶绿素含量[J].光谱学与光谱分析,2013,33(8):2203-2206.
[33]梁栋,管青松,黄文江,等.基于支持向量机回归的冬小麦叶面积指数遥感反演[J].农业工程学报,2013,29(7):117-123.
[34]Cutler D R,Edwards T C,Beard K H,et al.Random forests for classification in ecology[J].Ecology,2007,88(11):2783.