基于分数阶微分的橡胶树割胶期氮素含量定量检测
|
摘要
【目的】检测橡胶树割胶期叶片的实际氮素含量,建立橡胶树叶片光谱诊断模型,为实现橡胶树叶片氮素含量的快速无损检测提供参考依据。【方法】使用FieldSpec 3光谱仪采集割胶期橡胶树叶片的光谱反射率,分别以其原始光谱(R)、倒数光谱(1/R)、对数光谱(log R)和对数倒数光谱(1/log R)作为光谱信息,采用分数阶微分进行处理,获得不同分数阶阶次下的光谱数据,并通过竞争性自适应重加权算法(CARS)选择变量及偏最小二乘回归法(PLSR)建立橡胶树氮素光谱诊断模型。【结果】采用分数阶对橡胶树叶片R、1/R、log R和1/log R建立模型的最优均方根误差(RMSE)分别为0.1376、0.1175、0.1263和0.1505,且使用1/R数据建立的0.6阶模型表现最优,相关系数为0.9273,RMSE为0.1175,决定系数为0.8551。与整数阶算法相比,分数阶模型具有更强的预测能力,表明分数阶能充分挖掘光谱信息的有效信息,有效提高橡胶叶片氮含量光谱诊断模型的预测精度,实现橡胶树叶片氮素含量快速无损检测。【结论】应用近红外光谱技术并结合分数阶微分算法可快速无损检测橡胶树叶片氮素含量,为生产上橡胶树的精准可变量施肥提供技术支持。
【Objective】The actual nitrogen content of rubber leaves during tapping period was detected,and the spectral diagnostic model of rubber leaves was established,which could provide reference for the rapid and non-destructive detection of nitrogen content of rubber leaves.【Method】Fieldspec 3 spectrometer was used to collect the spectral reflectance of rubber leaves during rubber cutting period. The original spectrum(R),reciprocal spectrum(1/R),logarithmic spectrum(log R)and logarithmic reciprocal spectrum(1/log R)were used as spectral information respectively. Fractional calculus was used to process the spectral data under different fractional orders. The diagnostic model of nitrogen spectrum in rubber tree was established by selecting variables using competitive adaptive reweighted sampling(CARS) and partial least square regression(PLSR)algorithm.【Result】The optimum root mean square error(RMSE)of fractional-order for original spectrum,reciprocal spectrum,logarithmic spectrum and logarithmic reciprocal spectrum were 0.1376,0.1175,0.1263 and 0.1505,respectively. And the 0.6-order model based on reciprocal spectral data performed the best,the correlation coefficient index was 0.9273,RMSE was 0.1175 and determination coefficient was 0.8551. Compared with the integer-order algorithm,the fractional-order model had better prediction ability. The results showed that the fractional order was able to dig the effective information of spectral information and effectively improve the prediction accuracy of the spectral diagnostic model of nitrogen content in rubber leaves and finally achieved the rapid and non-destructive detection of nitrogen content of rubber leaves.【Conclusion】The application of near infrared spectroscopy and fractional calculus algorithm can quickly and non-destructively detect the nitrogen content in rubber leaves,providing technical support for precise variable fertilization of rubber trees in production.
【Objective】The actual nitrogen content of rubber leaves during tapping period was detected,and the spectral diagnostic model of rubber leaves was established,which could provide reference for the rapid and non-destructive detection of nitrogen content of rubber leaves.【Method】Fieldspec 3 spectrometer was used to collect the spectral reflectance of rubber leaves during rubber cutting period. The original spectrum(R),reciprocal spectrum(1/R),logarithmic spectrum(log R)and logarithmic reciprocal spectrum(1/log R)were used as spectral information respectively. Fractional calculus was used to process the spectral data under different fractional orders. The diagnostic model of nitrogen spectrum in rubber tree was established by selecting variables using competitive adaptive reweighted sampling(CARS) and partial least square regression(PLSR)algorithm.【Result】The optimum root mean square error(RMSE)of fractional-order for original spectrum,reciprocal spectrum,logarithmic spectrum and logarithmic reciprocal spectrum were 0.1376,0.1175,0.1263 and 0.1505,respectively. And the 0.6-order model based on reciprocal spectral data performed the best,the correlation coefficient index was 0.9273,RMSE was 0.1175 and determination coefficient was 0.8551. Compared with the integer-order algorithm,the fractional-order model had better prediction ability. The results showed that the fractional order was able to dig the effective information of spectral information and effectively improve the prediction accuracy of the spectral diagnostic model of nitrogen content in rubber leaves and finally achieved the rapid and non-destructive detection of nitrogen content of rubber leaves.【Conclusion】The application of near infrared spectroscopy and fractional calculus algorithm can quickly and non-destructively detect the nitrogen content in rubber leaves,providing technical support for precise variable fertilization of rubber trees in production.
引文
陈凯.2018.基于光谱技术的橡胶树氮素水平无损诊断策略研究[D].海口:海南大学.[Chen K.2018.Spectral diagnosis of nitrogen content level in rubber tree under naturally state of growth[D].Haikou:Hainan University.]
陈向阳,谭礼健.2017.基于自适应分数阶微分的医学图像增强算法[J].计算机应用研究,34(12):3895-3898.[Chen X Y,Tan L J.2017.Medical image enhancement algorithm based on adaptive fractional order differentiation[J].Application Research of Computer,34(12):3895-3898.]
陈贻钊,林清火,华元刚,罗微,林钊沐.2010.橡胶树叶片高光谱特征分析[J].热带亚热带植物学报,18(3):293-297.[Chen Y Z,Lin Q H,Hua Y G,Luo W,Lin Z M.2010.The analysis of hyperspectral characteristics of rubber leaves[J].Journal of Tropical and Subtropical Botany,18(3):293-297.]
郭伟,倪家健,李涛,邓玲.2011.基于时域的分数阶PID预测函数励磁控制器[J].仪器仪表学报,32(11):2461-2467.[Guo W,Ni J J,Li T,Deng L.2011.Improved fractionalorder PID predictive function excitation controller based on time domain[J].Chinese Journal of Scientific Instrument,32(11):2461-2467.]
蒋灿辰,唐荣年.2017.基于近红外光谱技术的成年橡胶树叶片氮素含量检测[J].安徽农业大学学报,44(3):429-433.[Jiang C C,Tang R N.2017.Detection of nitrogen in mature rubber tree leaves using near infrared spectroscopy[J].Journal of Anhui Agricultural University,44(3):429-433.]
李岚涛,汪善勤,任涛,马驿,魏全全,高雯晗,鲁剑巍.2016.基于高光谱的冬油菜叶片磷含量诊断模型[J].农业工程学报,32(14):209-218.[Li L T,Wang S Q,Ren T,Ma Y,Wei Q Q,Gao W H,Lu J W.2016.Evaluating models of leaf pho sphorus content of winter oilseed rape based on hyperspectral data[J].Transactions of the Chinese Society of Agricultural Engineering,32(14):209-218.]
芦鑫,高锦鸿,曹宇锋,张丽霞,黄纪念.2018.谷氨酰胺转氨酶改性花生蛋白质研究[J].河南农业科学,47(8):142-148.[Lu X,Gao J H,Cao Y F,Zhang L X,Huang J N.2018.Modification of peanut protein by transglutaminase[J].Journal of Henan Agricultural Sciences,47(8):142-148.]
纳嵘,胡波,史蓓.2018.近红外光谱技术测定苜蓿蛋白质含量的研究[J].畜牧与饲料科学,39(8):31-34.[Na R,Hu B,Shi B.2018.Measurement of protein content of alfalfa using near-infrared reflectance spectroscopy[J].Animal Husbandry and Feed Science,39(8):31-34.]
仇逊超.2018.红松仁脂肪的近红外光谱定量检测[J].江苏农业学报,34(3):692-698.[Qiu X C.2018.Quantitative detection of fat in peeled korean pine seeds using near infrared spectroscopy[J].Jiangsu Journal of Agricultural Sciences,34(3):692-698.]
吴继忠,王新宇,蓝黄博恩,谭浩东,殷思佳,陈晓宏,欧阳琴.2018.茶叶干燥过程水分含量的近红外光谱检测方法研究[J].农产品加工,(9):44-47.[Wu J Z,Wang X Y,Lanhuang B E,Tan H D,Yin S J,Chen X H,Ouyang Q.2018.Study on estimating the water content of tea in drying process using near infrared spectroscopy technique[J].Farm Products Processing,(9):44-47.]
吴伟斌,刘文超,李泽艺,张震邦,陈理,宋倩.2018.基于高光谱的茶叶含水量检测模型建立与试验研究[J].河南农业大学学报,52(5):818-824.[Wu W B,Liu W C,Li ZY,Zhang Z B,Chen L,Song Q.2018.Study on detection model establishment and experiment of tea water content based on hyperspectral[J].Journal of Henan Agricultural University,52(5):818-824.]
肖春华,吕银亮,刁万英,李少昆,王克如,陈兵,王琼.2015.基于冠层光谱角算法的小麦氮素营养监测[J].干旱地区农业研究,33(4):94-97.[Xiao C H,LüY L,Diao WY,Li S K,Wang K R,Chen B,Wang Q.2015.Monitoring nitrogen content of wheat on basis of an algorithm on the spectral angle of canopy[J].Agricultural Research in the Arid Areas,33(4):94-97.]
曾贞,孙恒,陈骏飞,徐天才,王宇萍,李坤,康平德,胡强.2018.FTIR结合化学计量学快速预测铁皮石斛中总多糖含量[J].江西农业学报,30(1):117-121.[Zeng Z,Sun H,Chen J F,Xu T C,Wang Y P,Li K,Kang P D,Hu Q.2018.Rapid forecast of total polysaccharide content in Dendrobium officinale by FTIR combined with chemometrics[J].Acta Agriculturae Jiangxi,30(1):117-121.]
张东,塔西甫拉提·特依拜,张飞,阿尔达克·克里木.2014.分数阶微分在盐渍土高光谱数据预处理中的应用[J].农业工程学报,30(24):151-160.[Zhang D,Tashpolat·Tiyip,Zhang F,Ardak·Kelimu.2014.Application of fractional differential in preprocessing hyperspectral data of saline soil[J].Transactions of the Chinese Society Agricultural Engineering,30(24):151-160.]
张晓艳,王丽丽,刘锋,封文杰,刘淑云,朱建华.2012.基于高光谱遥感的花生叶片氮积累量监测模型的研究[J].山东农业科学,44(3):7-12.[Zhang X Y,Wang L L,Liu F,Feng W J,Liu S Y,Zhu J H.2012.Study on monitoring model for nitrogen accumulation in peanut(Arachis hypogasa L.)leaves based on hyper-spectral remote sensing[J].Shandong Agricultural Sciences,44(3):7-12.]
赵永玲,侯之超.2013.基于分数导数的橡胶材料两种粘弹性本构模型[J].清华大学学报(自然科学版),53(3):378-383.[Zhao Y L,Hou Z C.2013.Two viscoelastic constitutive models of rubber materials using fractional derivations[J].Journal of Tsinghua University(Science and Technology),53(3):378-383.]
Ghamisi P,Couceiro M S,Benediktsson J A,Ferreira N M F.2012.An efficient method for segmentation of images based on fractional calculus and natural selection[J].Expert Systems with Applications,39(16):12407-12417.
Gianquinto G,Orsini F,Fecondini M,Mezzetti M,Sambo P,Bona S.2011.A methodological approach for defining spectral indices for assessing tomato nitrogen status and yield[J].European Journal of Agronomy,35(3):135-143.
Li F,Miao Y X,Feng G H,Yuan F,Yue S C,Gao X W,Liu Y Q,Liu B,Ustin S L,Chen X P.2014.Improving estimation of summer maize nitrogen status with red edgebased spectral vegetation indices[J].Field Crops Research,157:111-123.
Solari F,Shanahan J,Ferguson R,Schepers J,Gitelson A.2008.Active sensor reflectance measurements of corn nitrogen status and yield potential[J].Agronomy Journal,100(3):571-579.
Westad F,Martens H.2000.Variable selection in near infrared spectroscopy based on significance testing in partial least squares regression[J].Journal of Near Infrared Spectroscopy,8(2):117-124.
陈向阳,谭礼健.2017.基于自适应分数阶微分的医学图像增强算法[J].计算机应用研究,34(12):3895-3898.[Chen X Y,Tan L J.2017.Medical image enhancement algorithm based on adaptive fractional order differentiation[J].Application Research of Computer,34(12):3895-3898.]
陈贻钊,林清火,华元刚,罗微,林钊沐.2010.橡胶树叶片高光谱特征分析[J].热带亚热带植物学报,18(3):293-297.[Chen Y Z,Lin Q H,Hua Y G,Luo W,Lin Z M.2010.The analysis of hyperspectral characteristics of rubber leaves[J].Journal of Tropical and Subtropical Botany,18(3):293-297.]
郭伟,倪家健,李涛,邓玲.2011.基于时域的分数阶PID预测函数励磁控制器[J].仪器仪表学报,32(11):2461-2467.[Guo W,Ni J J,Li T,Deng L.2011.Improved fractionalorder PID predictive function excitation controller based on time domain[J].Chinese Journal of Scientific Instrument,32(11):2461-2467.]
蒋灿辰,唐荣年.2017.基于近红外光谱技术的成年橡胶树叶片氮素含量检测[J].安徽农业大学学报,44(3):429-433.[Jiang C C,Tang R N.2017.Detection of nitrogen in mature rubber tree leaves using near infrared spectroscopy[J].Journal of Anhui Agricultural University,44(3):429-433.]
李岚涛,汪善勤,任涛,马驿,魏全全,高雯晗,鲁剑巍.2016.基于高光谱的冬油菜叶片磷含量诊断模型[J].农业工程学报,32(14):209-218.[Li L T,Wang S Q,Ren T,Ma Y,Wei Q Q,Gao W H,Lu J W.2016.Evaluating models of leaf pho sphorus content of winter oilseed rape based on hyperspectral data[J].Transactions of the Chinese Society of Agricultural Engineering,32(14):209-218.]
芦鑫,高锦鸿,曹宇锋,张丽霞,黄纪念.2018.谷氨酰胺转氨酶改性花生蛋白质研究[J].河南农业科学,47(8):142-148.[Lu X,Gao J H,Cao Y F,Zhang L X,Huang J N.2018.Modification of peanut protein by transglutaminase[J].Journal of Henan Agricultural Sciences,47(8):142-148.]
纳嵘,胡波,史蓓.2018.近红外光谱技术测定苜蓿蛋白质含量的研究[J].畜牧与饲料科学,39(8):31-34.[Na R,Hu B,Shi B.2018.Measurement of protein content of alfalfa using near-infrared reflectance spectroscopy[J].Animal Husbandry and Feed Science,39(8):31-34.]
仇逊超.2018.红松仁脂肪的近红外光谱定量检测[J].江苏农业学报,34(3):692-698.[Qiu X C.2018.Quantitative detection of fat in peeled korean pine seeds using near infrared spectroscopy[J].Jiangsu Journal of Agricultural Sciences,34(3):692-698.]
吴继忠,王新宇,蓝黄博恩,谭浩东,殷思佳,陈晓宏,欧阳琴.2018.茶叶干燥过程水分含量的近红外光谱检测方法研究[J].农产品加工,(9):44-47.[Wu J Z,Wang X Y,Lanhuang B E,Tan H D,Yin S J,Chen X H,Ouyang Q.2018.Study on estimating the water content of tea in drying process using near infrared spectroscopy technique[J].Farm Products Processing,(9):44-47.]
吴伟斌,刘文超,李泽艺,张震邦,陈理,宋倩.2018.基于高光谱的茶叶含水量检测模型建立与试验研究[J].河南农业大学学报,52(5):818-824.[Wu W B,Liu W C,Li ZY,Zhang Z B,Chen L,Song Q.2018.Study on detection model establishment and experiment of tea water content based on hyperspectral[J].Journal of Henan Agricultural University,52(5):818-824.]
肖春华,吕银亮,刁万英,李少昆,王克如,陈兵,王琼.2015.基于冠层光谱角算法的小麦氮素营养监测[J].干旱地区农业研究,33(4):94-97.[Xiao C H,LüY L,Diao WY,Li S K,Wang K R,Chen B,Wang Q.2015.Monitoring nitrogen content of wheat on basis of an algorithm on the spectral angle of canopy[J].Agricultural Research in the Arid Areas,33(4):94-97.]
曾贞,孙恒,陈骏飞,徐天才,王宇萍,李坤,康平德,胡强.2018.FTIR结合化学计量学快速预测铁皮石斛中总多糖含量[J].江西农业学报,30(1):117-121.[Zeng Z,Sun H,Chen J F,Xu T C,Wang Y P,Li K,Kang P D,Hu Q.2018.Rapid forecast of total polysaccharide content in Dendrobium officinale by FTIR combined with chemometrics[J].Acta Agriculturae Jiangxi,30(1):117-121.]
张东,塔西甫拉提·特依拜,张飞,阿尔达克·克里木.2014.分数阶微分在盐渍土高光谱数据预处理中的应用[J].农业工程学报,30(24):151-160.[Zhang D,Tashpolat·Tiyip,Zhang F,Ardak·Kelimu.2014.Application of fractional differential in preprocessing hyperspectral data of saline soil[J].Transactions of the Chinese Society Agricultural Engineering,30(24):151-160.]
张晓艳,王丽丽,刘锋,封文杰,刘淑云,朱建华.2012.基于高光谱遥感的花生叶片氮积累量监测模型的研究[J].山东农业科学,44(3):7-12.[Zhang X Y,Wang L L,Liu F,Feng W J,Liu S Y,Zhu J H.2012.Study on monitoring model for nitrogen accumulation in peanut(Arachis hypogasa L.)leaves based on hyper-spectral remote sensing[J].Shandong Agricultural Sciences,44(3):7-12.]
赵永玲,侯之超.2013.基于分数导数的橡胶材料两种粘弹性本构模型[J].清华大学学报(自然科学版),53(3):378-383.[Zhao Y L,Hou Z C.2013.Two viscoelastic constitutive models of rubber materials using fractional derivations[J].Journal of Tsinghua University(Science and Technology),53(3):378-383.]
Ghamisi P,Couceiro M S,Benediktsson J A,Ferreira N M F.2012.An efficient method for segmentation of images based on fractional calculus and natural selection[J].Expert Systems with Applications,39(16):12407-12417.
Gianquinto G,Orsini F,Fecondini M,Mezzetti M,Sambo P,Bona S.2011.A methodological approach for defining spectral indices for assessing tomato nitrogen status and yield[J].European Journal of Agronomy,35(3):135-143.
Li F,Miao Y X,Feng G H,Yuan F,Yue S C,Gao X W,Liu Y Q,Liu B,Ustin S L,Chen X P.2014.Improving estimation of summer maize nitrogen status with red edgebased spectral vegetation indices[J].Field Crops Research,157:111-123.
Solari F,Shanahan J,Ferguson R,Schepers J,Gitelson A.2008.Active sensor reflectance measurements of corn nitrogen status and yield potential[J].Agronomy Journal,100(3):571-579.
Westad F,Martens H.2000.Variable selection in near infrared spectroscopy based on significance testing in partial least squares regression[J].Journal of Near Infrared Spectroscopy,8(2):117-124.