基于高光谱的玉米叶片氮含量监测模型
|
摘要
通过实测获取不同氮素营养水平下玉米冠层叶片光谱特性数据,利用可见光-近红外区域标准反射曲线的斜率和夹角等新型特征参数,结合预测学中最优权重组合原理,构建玉米叶片氮含量(LNC)监测模型。结果表明:光谱曲线夹角参数和斜率参数均与LNC存在较好的相关性,其中夹角参数Aγ和Aδ以及斜率参数Kr、Kb和Kn1的相关系数均在0.7左右;在单项监测模型中M(Kr/Kb)和M(Kb)以及M(Aδ/Aβ)和M(Aδ/Aα)的模型效果最好;最优模型由M(Kn1)和M(Kr/Kb)2个单项模型组合而成,其权重分别为0.245 5和0.754 5,最优模型的监测结果的决定系数(R2)为0.752 7,均方根误差(RMSE)为0.534,监测精度较单项模型明显提高。
The spectral features data of leaves in maize canopy under different level of nitrogen nutrition was obtained through actual measurement.The new spectral features such as slopes and angles extracted from the normalized reflectance curves in Visible-Near Infrared region was used and the optimal combination principle of the statistical forecasting were combined to build monitoring models for leaf nitrogen concentration of maize.The results showed that the spectral features such as slopes and angles had good correlation with LNC.The correlation coefficient of angles parameters(Aγ,and Aγ)and slopes parameters(Kr,Kband Kn1)was about 0.7.The models of M(Kr/Kb),M(Kb),M(Aδ/Aβ)and M(Aδ/Aα)were best in single monitoring model.The optimal model consisted of the models of M(Kn1)and M(Kr/Kb).Their weights were 0.245 5 and 0.754 5 respectively.The coefficient of determination(R2)was 0.752 7 for the monitoring results of the optimal model,and its root mean square error(RMSE)was 0.534.The monitoring precision of the optimal model was more improved than single models.
The spectral features data of leaves in maize canopy under different level of nitrogen nutrition was obtained through actual measurement.The new spectral features such as slopes and angles extracted from the normalized reflectance curves in Visible-Near Infrared region was used and the optimal combination principle of the statistical forecasting were combined to build monitoring models for leaf nitrogen concentration of maize.The results showed that the spectral features such as slopes and angles had good correlation with LNC.The correlation coefficient of angles parameters(Aγ,and Aγ)and slopes parameters(Kr,Kband Kn1)was about 0.7.The models of M(Kr/Kb),M(Kb),M(Aδ/Aβ)and M(Aδ/Aα)were best in single monitoring model.The optimal model consisted of the models of M(Kn1)and M(Kr/Kb).Their weights were 0.245 5 and 0.754 5 respectively.The coefficient of determination(R2)was 0.752 7 for the monitoring results of the optimal model,and its root mean square error(RMSE)was 0.534.The monitoring precision of the optimal model was more improved than single models.
引文
[1]WU C Y,NIU Z,TANG Q,et al.Estimating chlorophyll content from hyperspectral vegetation indices:modeling and validation[J].Agricultural and Forest Meteorology,2008,148(8/9):1230-1241.
[2]PENG Y,GITELSON A A.Remote estimation of gross primary productivity in soybean and maize based on total crop chlorophyll content[J].Remote Sensing of Environment,2012,117(2):440-448.
[3]ONOYAMA H,RYU C,SUGURI M,et al.Nitrogen prediction model of rice plant at panicle initiation stage using ground-based hyperspectral imaging:growing degree-days integrated model[J].Precison Agriculture,2015,16(3):558-570.
[4]王仁红,宋晓宇,李振海,等.基于高光谱的冬小麦氮素营养指数估测[J].农业工程学报,2014,30(19):191-198.
[5]李丹,李斐,胡云才,等.基于光谱指数波段优化算法的小麦玉米冠层含氮量估测[J].光谱学与光谱分析,2016,36(4):1150-1157.
[6]金梁,胡克林,田明明,等.夏玉米叶片分层氮素营养的高光谱诊断[J].光谱学与光谱分析,2013,33(4):1032-1037.
[7]刘冰峰,李军,赵刚峰,等.夏玉米叶片全氮含量高光谱遥感估算模型研究[J].植物营养与肥料学报,2012,18(4):813-824.
[8]周丽丽,冯汉宇,阎忠敏,等.玉米叶片氮含量的高光谱估算及其品种差异[J].农业工程学报,2010,26(8):195-199.
[9]CHEN P F,DRISS H,NICOLAS T,et al.New spectral indicator assessing the efficiency of crop nitrogen tretment in corn and wheat[J].Remote Sensing of Environment,2010,114(9):1987-1997.
[10]徐新刚,赵春江,王纪华,等.基于可见光-近红外新光谱特征和最优组合原理的大麦叶片氮含量监测[J].红外与毫米波学报,2013,32(4):351-358.
[11]李利,陈俊辉,樊明寿.使用叶绿素仪进行马铃薯氮素营养诊断的样本数确定[J].作物杂志,2015(4):135-137.
[12]夏文豪,刘涛,关钰,等.硝酸盐反射仪和SPAD法对玉米氮素营养诊断的比较[J].中国生态农业学报,2016,24(10):1339-1346.
[13]王福民,黄敬峰,徐俊锋,等.基于光谱波段自相关的水稻信息提取波段选择[J].光谱学与光谱分析,2008,28(5):1098-1101.
[14]李国强,吴士文,郑国清,等.基于冠层反射光谱的夏玉米叶片氮积累量估测[J].中国农学通报,2014,30(3):85-90.
[2]PENG Y,GITELSON A A.Remote estimation of gross primary productivity in soybean and maize based on total crop chlorophyll content[J].Remote Sensing of Environment,2012,117(2):440-448.
[3]ONOYAMA H,RYU C,SUGURI M,et al.Nitrogen prediction model of rice plant at panicle initiation stage using ground-based hyperspectral imaging:growing degree-days integrated model[J].Precison Agriculture,2015,16(3):558-570.
[4]王仁红,宋晓宇,李振海,等.基于高光谱的冬小麦氮素营养指数估测[J].农业工程学报,2014,30(19):191-198.
[5]李丹,李斐,胡云才,等.基于光谱指数波段优化算法的小麦玉米冠层含氮量估测[J].光谱学与光谱分析,2016,36(4):1150-1157.
[6]金梁,胡克林,田明明,等.夏玉米叶片分层氮素营养的高光谱诊断[J].光谱学与光谱分析,2013,33(4):1032-1037.
[7]刘冰峰,李军,赵刚峰,等.夏玉米叶片全氮含量高光谱遥感估算模型研究[J].植物营养与肥料学报,2012,18(4):813-824.
[8]周丽丽,冯汉宇,阎忠敏,等.玉米叶片氮含量的高光谱估算及其品种差异[J].农业工程学报,2010,26(8):195-199.
[9]CHEN P F,DRISS H,NICOLAS T,et al.New spectral indicator assessing the efficiency of crop nitrogen tretment in corn and wheat[J].Remote Sensing of Environment,2010,114(9):1987-1997.
[10]徐新刚,赵春江,王纪华,等.基于可见光-近红外新光谱特征和最优组合原理的大麦叶片氮含量监测[J].红外与毫米波学报,2013,32(4):351-358.
[11]李利,陈俊辉,樊明寿.使用叶绿素仪进行马铃薯氮素营养诊断的样本数确定[J].作物杂志,2015(4):135-137.
[12]夏文豪,刘涛,关钰,等.硝酸盐反射仪和SPAD法对玉米氮素营养诊断的比较[J].中国生态农业学报,2016,24(10):1339-1346.
[13]王福民,黄敬峰,徐俊锋,等.基于光谱波段自相关的水稻信息提取波段选择[J].光谱学与光谱分析,2008,28(5):1098-1101.
[14]李国强,吴士文,郑国清,等.基于冠层反射光谱的夏玉米叶片氮积累量估测[J].中国农学通报,2014,30(3):85-90.