周敏姑,邵国敏,张立元,姚小敏,韩文霆
无人机多光谱遥感反演冬小麦SPAD值
周敏姑1,邵国敏2,张立元2,姚小敏2,韩文霆3※
(1. 西北农林科技大学旱区节水农业研究院,杨凌 712100;2. 西北农林科技大学机械与电子工程学院,杨凌 712100;3. 西北农林科技大学水土保持研究所,杨凌 712100)
为研究无人机多光谱遥感5个波段光谱反射率反演冬小麦SPAD(Soil and Plant Analyzer Development)值的可行性,该研究采用六旋翼无人机搭载五波段多光谱相机,采集冬小麦拔节期、孕穗期、抽穗期、开花期的冠层光谱影像并提取反射率特征参数,建立SPAD值的反演模型。结果表明,当波长范围在蓝光、绿光和红光波段,冬小麦拔节期、孕穗期和开花期的无人机多光谱影像反射率参数与SPAD值呈负相关关系,而在抽穗期,二者呈正相关;当波长范围为红边及近红外波段,在整个生长期,二者均呈现正相关关系。该研究构建冬小麦SPAD值反演模型采用了主成分回归、逐步回归和岭回归法,经对比发现基于逐步回归法构建的模型效果最优,该模型的校正决定系数为0.77,主成分回归法次之,岭回归法较差。此外,冬小麦抽穗期多光谱反射率反演SPAD值效果最显著,主成分回归、岭回归和逐步回归3种回归模型的校正决定系数分别为0.72、0.74和0.77。该研究可为无人机多光谱遥感监测作物长势、实现精准农业生产管理提供技术依据。
无人机;遥感;冬小麦;多光谱影像;回归模型;SPAD
叶绿素含量SPAD(Soil and Plant Analyzer Development)值是农作物生长过程中重要的生化参数之一[1],对其含量的监测有助于衡量作物光合能力和生理损伤状况[2],从而有效评估作物的生长环境及水肥管理情况,快速、准确地获取农作物SPAD值是智慧农业发展的必要条件。
目前,农作物SPAD值的监测方法主要包括人工测量法和遥感监测法[3]。人工测量法以手持式叶绿素仪应用最多,但测量时和叶片接触面积仅有0.000 006 m2,必须进行大量反复测定才能降低测定值的变异,因此存在测量面积小,工作量大,数据代表性差等缺点[4],远不能满足作物的大面积精准化管理需求,所以利用遥感技术对作物进行大面积、快速、动态的无损监测被广泛研究,但遥感监测法由于其非接触的,远距离的探测特点,在精度方面不及人工测量法,因其能快速、动态、及时地获取田间数据,被研究人员应用于小麦的SPAD值监测[5]。目前最常用的方法为卫星遥感和地面遥感,王丽爱等[6]利用环境减灾卫星(HJ-1)遥感技术分析了2010—2013年江苏地区稻茬小麦不同生育期叶片SPAD值与8种植被指数的相关关系,建立的回归方程能够较好地估算SPAD值;李粉玲等[7]模拟高分一号(GF-1)卫星光谱反射率,研究了冬小麦SPAD值与18种宽波段光谱指数的关系,证明了基于绿色归一化植被指数、绿色比值植被指数和三角绿度指数等建立的冬小麦SPAD值估算模型效果较优;张锐等[8]利用便携式地物光谱仪研究了湖南地区油菜的冠层高光谱反射率和SPAD值的关系,建立SPAD值预测模型,得出基于支持向量机的预测模型反演精度最高(决定系数为0.913)的结论;殷紫等[9]基于地面高光谱测量技术在西北地区利用光谱参数红边面积与黄边面积的比值与油菜叶片SPAD值构建了能较好估算油菜SPAD值的反演模型(决定系数为0.79)。孙红等[5]利用便携式地物光谱仪研究了北京市昌平区冬小麦的5个生长期冠层光谱反射率和叶绿素含量的变化特征,并对二者的相关性进行了研究,分别建立了拔节期和孕穗期叶绿素含量线性监测模型。上述研究方法中,卫星遥感虽然能够实现对农作物SPAD值的大面积快速无损监测,但存在成本高、周期长、分辨率较低等缺点[10];而地面光谱仪扫描范围小,不易操作,且结果易受人为因素和周围环境影响[11-12]。无人机遥感平台既克服了地物光谱仪的工作量大、数据代表性差的缺点,又具有成本低、时效性强、分辨率高的优点,弥补了卫星遥感和地面遥感的不足[13-14],目前已有研究将无人机遥感技术应用于作物的生长参数监测,周敏姑等[15]基于无人机多光谱遥感技术构建7种植被指数,对杨凌地区冬小麦拔节后至孕穗前的生长阶段叶片的叶绿素含量进行反演,证明调整土壤亮度植被指数构建的一元二次线性回归模型反演精度最高(决定系数为0.866);魏青等[16]基于无人机多光谱遥感技术对北京市大兴区的冬小麦在不同施氮水平下冠层叶绿素含量进行监测,选取拔节期、抽穗期和灌浆期3个生育期的16种植被指数,采用2种回归分析方法建立了不同施氮水平下冬小麦冠层叶绿素含量估算模型。以上研究均是利用无人机多光谱影像构建常用植被指数或利用便携式地物光谱仪获取冬小麦冠层叶片光谱反射数据对冬小麦叶绿素含量进行估算,但利用无人机遥感的多光谱反射率因素对冠层叶片SPAD值的研究还鲜有报道。
综上,本研究采用无人机遥感技术结合地面监测的方法,选取冬小麦拔节期、孕穗期、抽穗期和开花期4个生长期,运用多光谱影像,研究不同波段反射率因素与冠层叶片SPAD值的关系,尝试对冠层叶片5个波段反射率与SPAD值之间建立不同的回归模型,并对模型进行精度评价,得出反演SPAD值的最佳回归方法和最佳生长期,以期为陕西关中地区冬小麦SPAD的遥感监测提供理论支持,并为农作物的长势监测、精准管理提供技术依据。
试验区位于西北农林科技大学旱区节水农业研究院,地处陕西关中平原中部的杨陵区(34°14′N~34°20′N,107°59′E~108°08′E),地势南低北高,海拔460 m,年降水量635.1~663.9 mm,年均气温12.9 ℃,属暖温带季风半湿润气候区,种植作物一年两熟,以冬小麦和夏玉米为主,当年10月中下旬进行冬小麦播种,次年6月初收获。
试验区东西向长度25 m,南北向长度162.5 m,行向由南到北,共划分为65个2.5 m×25 m的长方形小区,每个小区内选择1个1 m×1 m的样本区,样本区分别位于长方形小区中心点或中心点两端水平方向8 m处,整体呈S形分布。
根据景毅刚等[17]在气候变暖对陕西冬小麦生育期的影响中,对1986年以来陕西冬小麦生长发育始期观测资料进行分析认为,陕西冬小麦拔节期出现在3月下旬末,抽穗期出现在4月下旬前期,开花期出现在4月下旬后期,因此本试验中无人机多光谱影像和地面数据采集时间选择为2018年4月1日、8日、16日和27日,分别对应冬小麦冠层叶片光谱变化较为明显的拔节期、孕穗期、抽穗期和开花期,每次进行SPAD值地面数据采集时同步获取无人机遥感数据。
1.2.1 无人机多光谱遥感图像获取
试验采用团队研发的六旋翼无人机,搭载五波段多光谱相机(MicaSense RedEdge-M,美国),组成无人机多光谱信息采集系统,无人机和相机信息参数如表1所示。数据采集选择天气晴朗无风的日期,采集时间为15:00—16:00,无人机飞行高度60 m,航速5 m/s,航向和旁向重叠度均为80%,地面分辨率为4 cm/pixel。获取无人机多光谱影像前,首先在飞行区域内布置漫反射板(反射效率58%,尺寸3 m×3 m,GroupVIII,美国),用于多光谱影像像元亮度值(Digital Number,DN)的标定。多光谱相机镜头垂直向下,采集5种不同波长范围内的小麦冠层多光谱影像,5种波段中心波长分别为475(蓝光波段)、560(绿光波段)、668(红光波段)、717(红边波段)和840(近红外波段)nm。
表1 无人机和相机主要参数
1.2.2 冬小麦4个生长期SPAD值测量
无人机影像采集当日,同步在地面利用手持式叶绿素仪(SPAD-502Plus,日本)测量65个样本的SPAD值。样本区内选取具有代表性的7株小麦植株,测量每株倒二叶的叶尖、叶中、叶基3个部位SPAD值,求得平均值作为该植株的SPAD值,7株小麦的平均值作为该样本的SPAD值。
本研究采用Pix4Dmapper软件对获取的无人机多光谱影像进行拼接及处理。首先利用对应地面控制点数据对多光谱影像进行校正,生成数字正射影像图(Digital Orthophoto Map,DOM);然后利用灰板对多光谱影像进行反射率校正,获取试验地反射率影像,以.TIF格式存储;最后采用ENVI 5.1软件平台裁剪得到4个生长期的单波段光谱反射率影像,提取本研究区的平均反射率作为样本在该波段的光谱反射率(图1)。
分别提取冬小麦4个生长期的无人机多光谱影像光谱反射率数据及与之对应的同步测量地面数据,构成样本数据集,每个波段均获得65组数据,随机选取70% 的样本数据(45组数据)作为建模集,采用不同的回归分析方法构建SPAD值反演模型,再利用其余30% 的样本数据(20组数据)作为验证集,评价该SPAD值反演模型。构建冬小麦SPAD值反演模型时,若自变量之间存在多重共线性问题[18]会降低模型检验可靠性,导致分析结果不稳定[19-20]。因此,本研究采用主成分回归法(Principle Component Regression,PCR)、逐步回归法(Stepwise Regression,SR)与岭回归法(Ridge Regression,RR)作为建模方法,消除多重共线性问题,并验证模型的可靠性及稳定性。主成分回归法将5个波段降维,利用某几个主要波段的线性组合解决共线性问题[21];逐步回归法用实测SPAD值与单波段反射率特征参数进行简单回归,逐步引入其余波段,剔除不显著波段,使模型中的波段既显著又无多重共线性问题[22];岭回归法是一种改良的最小二乘估计法,以损失部分信息和降低精度为代价获得更可靠回归系数[23]。
图1 冬小麦4个生长期多光谱反射率影像
本研究选用决定系数(coefficient of determination,2)、均方根误差(Root Mean Squared Error,RMSE)综合评价冬小麦SPAD值反演模型精度[24]。在多元回归分析中,当回归模型增加一个解释变量,决定系数2会相应增大,即2是回归模型解释变量个数的非减函数,因此使用2来判断具有相同被解释变量和不同个数解释变量的回归模型优劣时存在不合理性。为了消除解释变量个数对决定系数2的影响,选择使用校正决定系数(adjusted coefficient of determination,2adj)对模型拟合效果进行评价。模型的2越接近1,相应的RMSE数值越小,则模型估算能力越好。R、2adj和RMSE的计算方法如式(1)~(3)所示
本研究首先对无人机多光谱影像进行拼接、裁剪等预处理,获得冬小麦4个生长期的单波段光谱影像,提取反射率特征参数,并分别建立5个波段反射率数据和实测SPAD值之间的相关关系;然后判断自变量之间的共线性问题,分别基于主成分回归、逐步回归和岭回归法构建SPAD值反演模型;最后对各个模型进行验证分析,优选冬小麦SPAD值的最佳反演模型。具体研究方案如图2所示。
图2 冬小麦SPAD值反演模型构建流程图
试验选取的65个样本实测SPAD值统计特征见表2,随着小麦生长期的推移,SPAD平均值整体呈上升趋势,此结果与王凯龙等[25]在干旱区冬小麦不同生长阶段的光谱特征与叶绿素含量估测研究中的结果一致。本研究中得到的SPAD值变异系数介于1%~10%之间,表现为弱变异[26]。
表2 冬小麦4个生长阶段SPAD值统计特征
本研究采用OriginLab软件,建立的冬小麦不同生长期光谱反射率与SPAD值的特征曲线如图3所示。冬小麦拔节期至开花期,冠层叶片光谱反射率与SPAD值的关系表现出相同的变化规律,蓝、红光波段的光辐射被叶片中的叶绿素吸收进行光合作用而形成2个低反射区,在绿光波段形成较小的反射峰,红边波段出现了高反射峰,在近红外波段均出现最强反射峰。这是由于小麦冠层在可见光区(400~700 nm)的反射率主要取决于叶绿素含量的多少,叶绿素含量多,吸收率高,反射率就低,蓝光波段和红光波段是植物叶绿素的显著吸收波段,在绿光区吸收较少故形成绿色反射峰,随着叶绿素含量的增加,红边位置反射率也增加,出现一个高反射峰,而近红外光谱区,光谱反射率一般受叶片内部细胞结构和的影响,叶绿素含量高的叶片,其内部细胞更为复杂,因而反射率高。对近红外区叶片光谱反射率和叶绿素含量的关系,已有学者进行过研究,武倩雯等[11]在基于近红外波段玉米叶绿素含量最佳预测模型研究中,为了探究近红外波段玉米光谱反射率与其叶绿素含量之间的关系,对玉米叶绿素含量与近红外光谱反射率及植被指数之间的关系进行分析,建立叶绿素含量最佳模型。结果表明,在近红外波段,光谱反射率与玉米叶绿素含量的相关性较大。近红外区叶片光谱反射率虽然影响因素较多,但此波段位于绿色植被强反射光谱区,其为叶片健康状况最灵敏的标志,对植物长势反映敏感,指示植物光合作用能否正常进行,因此近红外区与叶片叶绿素含量关系密切。
由图3可知,冬小麦从拔节期到开花期,冠层光谱反射率在可见光区随着SPAD值增大,反射率减小,至孕穗期达到最小,抽穗期开始增大,至开花期达到最大。红边和近红外波段,冠层光谱反射率从拔节期到开花期一直呈现上升趋势。出现该趋势的原因在于,小麦植株处于生长阶段,SPAD值逐渐增大,光合能力不断增强,叶绿素含量逐渐增加,叶片的绿色加深,对可见光吸收增加,反射率减小,另外拔节期到孕穗期叶片对地面未全覆盖,裸露的土壤会增强对可见光的吸收,导致无人机多光谱影像反射率降低;小麦抽穗期阶段,由于植株冠层变黄,对可见光反射增强,吸收作用减弱,冠层光谱反射率开始增加,开花期达到最大。
图3 冬小麦4个生长阶段叶片光谱反射率随SPAD值变化特征
本研究首先采用统计分析软件SPSS 22.0分析5个波段叶片反射率和其SPAD值之间的相关关系(如表3所示)。由表3可知,当多光谱相机波长范围处于蓝光、绿光和红光波段时,冬小麦拔节期、孕穗期和开花期的无人机多光谱图像反射率参数与SPAD值呈负相关关系,而在抽穗期,二者呈正相关;当波长范围为红边及近红外波段,二者在整个生长期均呈现正相关关系。
此外,不同生长期的小麦叶片反射率与SPAD值的相关程度不同。开花期绿光波段小麦叶片反射率与SPAD值相关系数绝对值最高,为0.89,而拔节期近红外波段小麦叶片反射率与SPAD值相关系数最小,为0.71。通常认为相关系数为0.5~0.8表现为显著相关,0.8~1.0表现为高度相关[25];单波段小麦叶片反射率与SPAD值的显著相关性说明了采用5个波段反射率参数建立SPAD值的估算模型的可行性。
在上述相关性分析的基础上,本研究对建模数据集(45个样本)波段反射率与SPAD值进行多元线性回归分析。在回归分析前,选用容忍度[27]对5个波段反射率之间进行共线性判断,结果如表4所示。容忍度的取值在(0,1)之间,值越小,则多重共线性越严重[28];通常认为容忍度<0.1时,存在严重的多重共线性问题[29]。
表3 冬小麦4个生长阶段SPAD值与单波段光谱反射率相关性分析
注:** 表示在0.01水平上显著相关。
Note: ** indicates correlation is significant at 0.01 level。
表4 冬小麦4个生长阶段叶片单波段光谱反射率间容忍度统计分析
由表4可知,拔节期波段1、3、4,孕穗期波段4和开花期波段1、4、5的容忍度>0.1,其余均<0.1。这表明5个波段之间存在较严重的多重共线性问题,因此本研究分别采用主成分回归、逐步回归和岭回归法构建SPAD值反演模型,各模型的评价指标如表5所示。
表5 冬小麦4个生长阶段光谱反射率与SPAD值回归分析结果
注:为SPAD预测值;1、2、3、4和5分别为蓝光、绿光、红光、红边和近红外波段的光谱反射率。共65个样本,建模样本45个。
Note:is the predicted SPAD values;1,2,3,4,and5is the spectral reflectance of blue, green, red, red-edge and near-infrared band, respectively. There are 65 samples, including 45 modeling samples.
由表5可知,采用3种回归分析法建立的冬小麦4个生长期的SPAD值反演模型中,表达模型的波段与波段数目均不相同,这说明在冬小麦生长的不同阶段,SPAD对波段光谱的敏感性不同,拔节期最敏感波段为近红外波段,孕穗期为蓝光波段,抽穗期近红外波段,开花期为绿光和红光波段。
此外,3种回归分析法建立的冬小麦SPAD值反演模型在小麦不同生长期的计算精度有所差异。其中,拔节期主成分回归模型的精度检验结果最优,而岭回归模型的精度略优于逐步回归模型,由主成分回归模型得到的叶绿素SPAD估算值与实测值之间的2adj为0.68,RMSE为0.58;孕穗期主成分回归模型各项检验指标精度仍然最优;而抽穗期逐步回归模型精度最优,其2adj为0.77,RMSE为0.61;开花期3种回归模型的2adj值较为接近,但逐步回归模型的RMSE较小,为0.63,表明其精度较高,抽穗期建立的SPAD值回归模型精度要高于其他生长期。
比较每个生长期筛选出的最优模型可以看出,抽穗期构建的逐步回归模型的2adj最高、RMSE最小,故冬小麦抽穗期建立的逐步回归模型精度优于其他模型,可作为冬小麦SPAD值反演的最佳模型(图4a)。为验证模型的可靠性,采用验证数据集(30%的样本数据,20个样本)进行验证,结果如图4b所示。结果表明,冬小麦SPAD的预测值与实测值拟合效果较好(2= 0.73,RMSE= 0.56,= 20)。因此,冬小麦抽穗期基于逐步回归法构建的模型能较好地反演SPAD值。
图4 基于逐步回归法的冬小麦抽穗期SPAD值反演模型模拟及预测值与实测值的关系
叶绿素相对含量SPAD值是农作物的主要生化参数之一,其含量变化与作物的生存状况、生长态势密切相关,快速、准确、动态地监测作物SPAD值,对智慧农业的发展具有重要意义[30]。本研究选取冬小麦拔节期、孕穗期、抽穗期和开花期4个生长期,利用无人机遥感平台获取多光谱影像,提取叶片光谱反射率数据构建冬小麦SPAD值的反演模型。结果发现,4个生长期构建的模型中,抽穗期建立的3种回归模型精度均高于其他生长期,其中逐步回归法构建的模型精度最高,2adj为 0.77,RMSE为0.61;李粉玲等[7]等基于高分一号卫星影像数据提取18种高光谱植被指数估算冬小麦叶片SPAD值,认为拔节期构建的模型效果最优。两种研究方法结果不同的原因可能是:采用植被指数研究SPAD值时,常用的植被指数均是通过蓝光、绿光、红光和近红外波段的光谱反射率经过波段运算获得的宽带绿度指数[31],忽略了红光波段与近红外区域的红边部分,红边是由于植被在红光波段叶绿素强烈的吸收与近红外波段光在叶片内部的多次散射而形成的强反射造成的[32],植被覆盖度越高,红边植被指数对SPAD值越敏感,当冬小麦进入抽穗期,植株由生殖生长转向营养生长,叶片、叶稍鲜重达到峰值,覆盖度属于整个生长期中最优时期[33],此阶段光谱反射率对红边波段敏感性较高,因此红边位置对研究作物SPAD值非常重要。王凯龙等[25]基于地面光谱分析仪提取15种高光谱植被指数估算冬小麦SPAD值,将红边内一阶微分最大值处的波长(Red Edge Position,REP)加入研究,结果认为开花期构建的模型效果最优。综上,对作物SPAD值进行研究时,获取光谱数据的方法、植被指数的类别、建模方法的选用等因素,导致不同研究方法得到的模型不同。因此,利用遥感技术对农作物生化参数进行监测时,光谱影像中获取信息参数的精度、植被指数的适用性以及建模方法的选用还需要进一步研究,光谱反射率与SPAD值之间的相关性高,也并不能说明该波长处的反射率就一定对叶绿素含量有指示作用,需要综合考虑冬小麦的群体特征叶面积指数、叶片内部结果、植被覆盖度以及土壤背景等因素的影响。
本研究利用无人机多光谱遥感技术结合地面监测数据研究了叶片光谱反射率参数反演冬小麦SPAD(Soil and Plant Analyzer Development)值的可行性,得出以下结论:
1)冬小麦从拔节期到开花期,冠层光谱反射率与SPAD值特征曲线分析结果表明,拔节期冠层光谱反射率在可见光区随着SPAD值增大,反射率减小,至孕穗期达到最小,抽穗期开始增大,至开花期达到最大。红边和近红外波段,冠层光谱反射率从拔节期到开花期一直呈现上升趋势。
2)通过对冬小麦SPAD值与不同波段的无人机多光谱影像反射率进行相关性分析得知,在蓝光、绿光和红光波段,光谱反射率和SPAD值在冬小麦拔节期、孕穗期和开花期均呈显著负相关关系,而在抽穗期呈正相关;在红边和近红外波段,SPAD值与光谱反射率在冬小麦4个生长期均呈现正相关关系;相关系数绝对值最大为0.89,最小为0.71。
3)利用3种回归法建立了基于5个波段叶片光谱反射率的SPAD值反演模型。经验证,冬小麦4个生长期中抽穗期建立的模型精度最高,为SPAD值的最佳反演阶段,其次是开花期、拔节期和孕穗期;3种回归方法建立的反演模型中,抽穗期基于逐步回归法反演效果最优,其决定系数为0.77,均方根误差为0.61。
4)通过对3种回归法构建的冬小麦SPAD值反演模型进行分析得知,不同生长阶段SPAD值对波段光谱的敏感性不同,拔节期最敏感波段为近红外波段,孕穗期为蓝光波段,抽穗期为近红外波段,开花期为绿光和红光波段。
上述结论表明利用无人机平台获取多波段光谱反射率,建立冬小麦SPAD值反演模型,具有较好的预测精度,研究结果可为作物SPAD值的遥感反演研究提供进一步参考,以期为精准农业的管理和决策奠定科学基础和提供技术支持。
[1] Zhang Suming, Zhao Gengxing, Lang Kun, et al. Integrated satellite, Unmanned Aerial Vehicle (UAV) and ground inversion of the SPAD of winter wheat in the reviving stage[J]. Sensors, 2019, 19(7): 1-17.
[2] 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. Liang Liang, Yang Minhua, Zhang Lianpeng, et al. Chlorophyll content inversion with hyperspectral technology for wheat canopy based on SVR algorithm[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012, 28(20): 162-171. (in Chinese with English abstract)
[3] 岳学军,全东平,洪添胜,等. 柑橘叶片叶绿素含量高光谱无损检测模型[J]. 农业工程学报,2015,31(1):294-302. Yue Xuejun, Quan Dongping, Hong Tiansheng, et al. Non-destructive hyperspectral measurement model of chlorophyll content for citrus leaves[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31(1): 294-302. (in Chinese with English abstract)
[4] 马明洋,许童羽,周云成,等. 东北粳稻叶绿素相对含量的无人机高清影像检测方法[J]. 沈阳农业大学学报,2017,48(6):757-762. Ma Mingyang, Xu Tongyu, Zhou Yuncheng, et al. UAV HD image detection method for SPAD in northeast japonica rice[J]. Journal of Shenyang Agricultural University, 2017, 48(6): 757-762. (in Chinese with English abstract)
[5] 孙红,李民赞,赵勇,等. 冬小麦生长期光谱变化特征与叶绿素含量监测研究[J]. 光谱学与光谱分析,2010,30(1):192-196. Sun Hong, Li Minzan, Zhao Yong, et al. The spectral characteristics and chlorophyll content at winter wheat growth stages[J]. Spectroscopy and Spectral Analysis, 2010, 30(1): 192-196. (in Chinese with English abstract)
[6] 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. Wang Liai, Ma Chang, Zhou Xudong, et al. Estimation of wheat leaf SPAD value using RF algorithmic model and remote sensing data[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(1): 259-265. (in Chinese with English abstract)
[7] 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. Li Fenling, Wang Li, Liu Jing, et al. Remote sensing estimation of SPAD value for wheat leaf based on GF-1date[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(9): 273-281. (in Chinese with English abstract)
[8] 张锐,廖桂平,王访,等. 基于冠层高光谱的油菜角果期红边参数及叶片SPAD值反演模型[J]. 江苏农业科学,2019,47(20):255-259. Zhang Rui, Liao Guiping, Wang Fang, et al. Red edge parameters and SPAD inversion model of rapeseed based on canopy hyperspectral data[J]. Jiangsu Agricultural Sciences, 2019, 47(20): 255-259. (in Chinese with English abstract)
[9] 殷紫,常庆瑞,刘淼,等. 基于光谱指数的不同生育期油菜叶片SPAD估测[J]. 西北农林科技大学学报:自然科学版,2017,45(5):66-72. Yin Zi, Chang Qinrui, Liu Miao, et al. Estimation of rape leaf SPAD in different periods based on spectral indices[J]. Journal of Northwest A&F University: Natural Science Edition, 2017, 45(5): 66-72. (in Chinese with English abstract)
[10] 陈鹏,冯海宽,李长春,等. 无人机影像光谱和纹理融合信息估算马铃薯叶片叶绿素含量[J]. 农业工程学报,2019,35(11):63-74. Cheng Peng, Feng Haikuan, Li Changchun, et al. Estimation of chlorophyll content in potato using fusion of texture and spectral features derived from UAV multispectral image[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(11): 63-74. (in Chinese with English abstract)
[11] 武倩雯,熊黑钢,王莉锋,等. 基于近红外波段玉米叶绿素含量最佳预测模型研究[J]. 中国农学通报,2015,31(15):260-264. Wu Qianwen, Xiong Heigang, Wang Fengli, et al. Optimum predection model of maize leaf chlorophyll content based on near-infrared band[J]. Chinese Agricultural Science Bulletin, 2015, 31(15): 260-264. (in Chinese with English abstract)
[12] 周雪,王芳,赵庚星. 基于小麦冠层近地多光谱图像的叶绿素(SPAD值)估测方法[J]. 山东农业科学,2016,48(6):138-141. Zhou Xue, Wang Fang, Zhao Gengxing. Estimation method of chlorophyll content (SPAD value) based on near ground multispectral images of wheat canopy[J]. Shandong Agricultural Sciences, 2016, 48(6): 138-141. (in Chinese with English abstract)
[13] 秦占飞,常庆瑞,谢宝妮,等. 基于无人机高光谱影像的引黄灌区水稻叶片全氮含量估测[J]. 农业工程学报,2016,32(23):77-85. Qin Zhanfei, Chang Qingrui, Xie Baoni, et al. Rice leaf nitrogen content estimation based on hyperspectral imagery of UAV in Yellow River diversion irrigation district[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(23): 77-85. (in Chinese with English abstract)
[14] 裴浩杰,冯海宽,李长春,等. 基于综合指标的冬小麦长势无人机遥感[J]. 农业工程学报,2017,33(20):74-82. Pei Haojie, Feng Haikuan, Li Changchun, et al. Remote sensing monitoring of winter wheat growth with UAV based on comprehensive index[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(20): 74-82. (in Chinese with English abstract)
[15] 周敏姑,邵国敏,张立元,等. 基于无人机遥感的冬小麦叶绿素含量多光谱反演[J]. 节水灌溉,2019(9):40-45. Zhou Mingu, Shao Guomin, Zhang Liyuan, et al. Multi-spectral inversion of SPAD value of winter wheat based on unmanned aerial vehicle remote sensing[J]. Water Saving Irrigation, 2019(9): 40-45. (in Chinese with English abstract)
[16] 魏青,张宝忠,魏征,等. 基于无人机多光谱遥感的冬小麦叶绿素含量估测研究[J]. 麦类作物学报,2020,40(3):365-372. Wei Qing, Zhang Baozhong, Wei Zheng, et al. Estimation of canopy chlorophyll content in winter wheat by UAV multispectral remote sensing[J]. Journal of Triticeae Crops, 2020, 40(3): 365-372. (in Chinese with English abstract)
[17] 景毅刚,范建忠,高茂盛. 气候变暖对陕西冬小麦生育期的影响[J]. 麦类作物学报,2013,33(2):389-396. Jing Yigang, Fan Jianzhong, Gao Maosheng. Influences of the climate warming on developmental stages of winter wheat in Shaanxi[J]. Journal of Triticeae Crops, 2013, 33(2): 389-396. (in Chinese with English abstract)
[18] Ullah M I, Aslam M, Altaf S, et al. Some new diagnostics of multicollinearity in linear regression model[J]. Sains Malaysiana, 2019, 48(9): 2051-2060.
[19] 田明璐,班松涛,常庆瑞,等. 基于无人机成像光谱仪数据的棉花叶绿素含量反演[J]. 农业机械学报2016,47(11):285-293. Tian Minglu, Ban Songtao, Chang Qingrui, et al. Estimation of SPAD value of cotton leaf using hyperspectral images from UAV-based imaging spectroradiometer[J]. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(11): 285-293. (in Chinese with English abstract)
[20] 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. Pan Bei, Zhao Gengxing, Zhu Xicun, et al. Estimation of chlorophyll content in apple tree canopy based on hyperspectral parameters[J]. Spectroscopy and Spectral Analysis, 2013, 33(8): 2203-2206. (in Chinese with English abstract)
[21] 卢二坡,张焕明. 基于稳健主成分回归的统计数据可靠性评估方法[J]. 统计研究,2011,28(8):21-27. Lu Erpo, Zhang Huanming. An evaluation method of reliability of published statistics data based on robust principal component regression method[J]. Statistical Research, 2011, 28(8): 21-27. (in Chinese with English abstract)
[22] 游士兵,严研. 逐步回归分析法及其应用[J]. 统计与决策,2017,7(14):31-35. You Shibing, Yan Yan. Stepwise regression analysis and its application[J]. Statistics & Decision, 2017, 7(14): 31-35. (in Chinese with English abstract)
[23] 开璇,张莉莉. 主成分回归和岭回归在新疆农业经济的应用[J]. 辽宁农业职业技术学院学报,2017,19(1):57-61. Kai Xuan, Zhang Lili. Application of Xinjiang agricultural economy based on ridge regression and principal component regression[J]. Journal of Liaoning Agricultural Technical College, 2017, 19(1): 57-61. (in Chinese with English abstract)
[24] Ma Xiaodan, Feng Jiarui, Guan Haiou, et al. Prediction of chlorophyll content in different light areas of apple tree canopies based on the color characteristics of 3D reconstruction[J]. Remote Sensing, 2018, 10(3): 1-16.
[25] 王凯龙,熊黑钢,张芳. 干旱区冬小麦不同生长阶段的光谱特征与叶绿素含量估测研究[J]. 干旱区资源与环境,2013,27(11):45-49. Wang Kailong, Xiong Heigang, Zhang Fang. The spectral characteristics and the best model for chlorophyll content at winter wheat growth stages in arid land[J]. Journal of Arid Land Resources and Environment, 2013, 27(11): 45-49. (in Chinese with English abstract)
[26] 王海峰,张智韬,Amon K,等. 基于灰度关联-岭回归的荒漠土壤有机质含量高光谱估算[J]. 农业工程学报,2018,34(14):124-131. Wang Haifeng, Zhang Zhitao, Amon K, et al. Hyperspectral estimation of desert soil organic matter content based on gray correlation-ridge regression model[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(14): 124-131. (in Chinese with English abstract)
[27] 毛博慧,孙民赞,孙红,等. 冬小麦苗期叶绿素含量检测光谱学参数寻优[J]. 农业工程学报,2017,33(2):164-169. Mao Bohui, Sun Minzan, Sun Hong, et al. Optimization of spectroscopy parameters and prediction of chlorophyll content at seeding stage of winter wheat[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(2): 164-169. (in Chinese with English abstract)
[28] Gamze G, Hatice S. Examination of dimension reduction performances of PLSR and PCR techniques in data with multicollinearity[J]. Iranian Journal of Science & Technology, 2019, 43(3): 969-978.
[29] 张智韬,王海峰,韩文霆,等. 基于无人机多光谱遥感的土壤含水率反演研究[J]. 农业机械学报,2018,49(2):173-181. Zhang Zhitao, Wang Haifeng, Han Wenting, et al. Inversion of soil moisture content based on multispectral remote sensing of UAV[J]. Transactions of the Chinese Society for Agricultural Machinery, 2018, 49(2): 173-181. (in Chinese with English abstract)
[30] Ahmed A D, Abdulah E K, Ismael B. et al. Solving multicollinearity problem of gross domestic product using ridge regression method[J]. Periodicals of Engineering and Natural Sciences, 2020, 2(8): 668-672.
[31] 郭燕,程永政,黎世民,等. 区域尺度冬小麦叶绿素含量的高光谱预测和空间变异研究[J]. 麦类作物学报,2017,37(7):970-977. Guo Yan, Cheng Yongzheng, Li Shimin, et al. Hyperspectral-based estimation and spatial variability of chlorophyll content of winter wheat in regional scale[J]. Journal of Triticeae Crops, 2017, 37(7): 970-977. (in Chinese with English abstract)
[32] 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. Luo Dan, Chang Qingui, Qi Yanbing, et al. Estimation model chlorophyll content in winter wheat canopy based on spectral indices[J]. Journal of Triticeae Crops, 2016, 36(9): 1225-1233. (in Chinese with English abstract)
[33] 鞠昌华,田永超,朱艳,等. 小麦叠加叶片的叶绿素含量光谱反演研究[J]. 麦类作物学报,2008,28(6):1068-1074. Ju Changhua, Tian Yongchao, Zhu Yan, et al. Spectral inverse study of stacked leaf chlorophyll concentration in wheat[J]. Journal of Triticeae Crops, 2008, 28(6): 1068-1074. (in Chinese with English abstract)
Inversion of SPAD value of winter wheat by multispectral remote sensing of unmanned aerial vehicles
Zhou Mingu1, Shao Guomin2, Zhang Liyuan2, Yao Xiaomin2, Han Wenting3※
(1.,,712100,; 2.,,712100,; 3.,,712100,)
Remote sensing technology has been widely used to monitor the changes in SPAD, which is an important parameter. In this study, the multispectral images were acquired by a six-rotor unmanned aerial vehicle, and the SPAD of winter wheat was measured to carry out the estimation research. The four growth stages with the most obvious changes in SPAD were selected, namely the jointing stage, booting stage, heading stage, and flowering stage. The camera with five bands (475, 560, 668, 717, and 840 nm) was used to collect multispectral canopy leaves at the four stages. A total of four data collections were performed to extract spectral reflectance data and the SPAD was measured from 1stApril to 27thApril 2018. A total of 65 samples were selected and recorded with GPS. The test area was divided into 65 sample zones with each one measuring 2.5 m×25 m, of which one sample area of 1 m × 1 m was selected. All the zones were in a rectangle, so they could be evenly distributed 8 m from the center of the cell in the horizontal direction. The overall samples were S-shaped distribution. The samples in the middle were located at the center of the rectangular cell. The SPAD of 65 samples were measured by SPAD-502 chlorophyll meter at the same time when the UAV data was collected. In the sample area, seven leaves of different canopy parts were selected to measure the tip, middle, and base. The average of the three parts was used as the SPAD values of the leaf. Finally, the average value of the leaf blades was taken as the final SPAD value of the sample. The canopy reflectance data was extracted from multispectral images. And then the correlation coefficients of SPAD values and spectral reflectance data in four growth stages were analyzed. Herein, the reflectivity of single-band and SPAD directly had serious collinearity problems so principal component regression, stepwise regression, and ridge regression these three methods were chosen to solve it. After that, the SPAD inversion models were established separately by using the reflectance data and the SPAD values as the data source. The best inversion model and stage were selected by comparison. The results showed that a high correlation was obtained between the SPAD and canopy spectral reflectance. In the visible light band, the negative correlation was observed between canopy spectral reflectance and SPAD at the jointing stage, booting stage, and flowering stage. On the contrary, it was a positive correlation at the heading stage and a positive correlation at the red-edge and near-infrared bands at all four stages. Compared with the main bands in the model expression, the frequency of passing the screening in different growth stages was different. The highest passing frequency was the near-infrared band in the jointing stage. The blue band was selected at the booting stage, the near-infrared band at the heading stage, and the green and red bands at the flowering stage. This study compared the prediction accuracy of the models established by three regression methods. The results showed that the models of stepwise regression established at the heading stage had the highest inversion accuracy with the adjusted coefficient of determination was 0.77, and the root mean square error was 0.61. The validation showed the coefficient of determination was 0.73, and the root mean square error was 0.56. It indicated that the model could be used to estimate the crop coefficient. Compared with the four periods, the heading stage was the best inversion stage of SPAD value. The study results proved the feasibility of inversion of the winter wheat SPAD value by unmanned aerial vehicle multispectral remote sensing, and at the same time, it could provide a reference for the rapid monitoring of the SPAD value of other crops.
unmanned aerial vehicle; remote sensing; winter wheat; multispectral image; regression model; SPAD
周敏姑,邵国敏,张立元,等. 无人机多光谱遥感反演冬小麦SPAD值[J]. 农业工程学报,2020,36(20):125-133.doi:10.11975/j.issn.1002-6819.2020.20.015 http://www.tcsae.org
Zhou Mingu, Shao Guomin, Zhang Liyuan, et al. Inversion of SPAD value of winter wheat by multispectral remote sensing of unmanned aerial vehicles[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2020, 36(20): 125-133. (in Chinese with English abstract) doi:10.11975/j.issn.1002-6819.2020.20.015 http://www.tcsae.org
2020-03-04
2020-05-25
“十三五”国家重点研发计划项目(2017YFC0403203);杨凌示范区产学研用协同创新重大项目(2018CXY-23)
周敏姑,实验师,主要从事农业智能检测、材料分析与检测研究。Email:zmingu@163.com
韩文霆,博士,研究员,主要从事无人机遥感与精准灌溉技术研究。Email:hanwt2000@126.com
10.11975/j.issn.1002-6819.2020.20.015
S252
A
1002-6819(2020)-20-0125-09