Conference PaperPDF Available

Abstract and Figures

In order to verify the possibility of in season nitrogen monitoring, using spectral canopy sensors in sugarcane fields, an randomized block design experiment with 4 treatment (N rates: 0, 60, 120 and 180 kg ha-1) and 4 replications was installed in a sugarcane-production Field in the municipality of Pradópolis, São Paulo state, Brazil. The FieldSpec 4 device was used to evaluate the canopy spectral response in 5 points previously marked within each plot, when the crop was on average height of 0.7 m in the main stem. In these collection points were sampled relative chlorophyll content (RCC) with SPAD-502 portable chlorophill meter, used as a reference for models development. The correlations of each band with N rates applied and the RCC were performed. The most significant spectral wavelength correlation to RCC were 550, 720 and 750 nm (r =-0.50,-0.69 and-0.77, respectively). Data from green (550 nm) and Near Infrared (750 nm) spectral region were used to the calculation of the Normalized Difference Vegetation Index (NDVI), showing a good correlation to RCC (r = 0.68). Linear combinations of selected bands and NDVI was performed by stepwise multiple linear regression, using 64 samples, where the equation parameters showed adjustment of R² = 0.69 and a standard error of 0.81 and in the validation phase the estimated and observed data, showed correlation of 0.80 with the trendline close to 1:1 line in the scatter plot. The results showed that the green (550 nm), red-edge (720 nm) and near infrared (750 nm) was the most significant spectral regions to estimate de RCC content in sugarcane.
Content may be subject to copyright.
Dados espectrais de dossel de cana-de-açúcar para predição do teor relativo de clorofila
Peterson Ricardo Fiorio 1
Juliano Araujo Martins 2
Pedro Paulo da Silva Barros 1
José Paulo Molin 1
Lucas Rios do Amaral 1
1 Universidade de São Paulo - USP/ESALQ
Avenida Pádua Dias 11, CEP - 13418-900 - Piracicaba - SP, Brasil
{fiorio, pedropaulo, jpmolin}@usp.br, lucasamaral@agronomo.eng.br
2 Instituto Federal de Mato Grosso - IFMT
Avenida Tancredo Neves 543 - 78890-000 - Sorriso - MT, Brasil
julianoaraujo3@gmail.com
Abstract. In order to verify the possibility of in season nitrogen monitoring, using spectral canopy sensors in
sugarcane fields, an randomized block design experiment with 4 treatment (N rates: 0, 60, 120 and 180 kg ha-1)
and 4 replications was installed in a sugarcane-production Field in the municipality of Pradópolis, São Paulo state,
Brazil. The FieldSpec 4 device was used to evaluate the canopy spectral response in 5 points previously marked
within each plot, when the crop was on average height of 0.7 m in the main stem. In these collection points were
sampled relative chlorophyll content (RCC) with SPAD-502 portable chlorophill meter, used as a reference for
models development. The correlations of each band with N rates applied and the RCC were performed. The most
significant spectral wavelength correlation to RCC were 550, 720 and 750 nm (r =-0.50, -0.69 and -0.77,
respectively). Data from green (550 nm) and Near Infrared (750 nm) spectral region were used to the calculation
of the Normalized Difference Vegetation Index (NDVI), showing a good correlation to RCC (r = 0.68). Linear
combinations of selected bands and NDVI was performed by stepwise multiple linear regression, using 64 samples,
where the equation parameters showed adjustment of R² = 0.69 and a standard error of 0.81 and in the validation
phase the estimated and observed data, showed correlation of 0.80 with the trendline close to 1:1 line in the scatter
plot. The results showed that the green (550 nm), red-edge (720 nm) and near infrared (750 nm) was the most
significant spectral regions to estimate de RCC content in sugarcane.
Palavras-chave: Nitrogen, In-season evaluation, Reflectance.
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6313
1. Introdução
A utilização de clorofilômetros para o monitoramento do estado nutricional do nitrogênio em
culturas, apesar de apresentar boa sensibilidade, possui grande limitação para utilização em
agricultura de precisão, uma vez que no monitoramento de grandes áreas, grande número de
amostras são requeridas para a geração de um mapa de variabilidade fidedigna a realidade de
campo (MIAO et al., 2008).
Neste caso, sensores de dossel acoplados em maquinário agrícola poderiam ser mais
eficientes, devido a sua capacidade de gerar mapas com grande detalhamento espacial, onde,
por meio da implementação de um algoritmo, permite a quantificação e aplicação de
fertilizantes nitrogenados em tempo real.
Estas técnicas já vêm sendo estudadas em muitas áreas produtivas do globo, em diferentes
culturas e sistemas produtivos, onde resultados promissores têm sido alcançados, tendo
destaque trabalhos com milho, trigo (CHEN et al., 2010), arroz (SONG et al., 2011). No entanto
estudos neste sentido na cultura da cana-de-açúcar ainda são escassos (MIPHOKASAP et al.,
2012; PORTZ et al., 2011).
Tendo em vista que a cana-de-açúcar possui grande importância econômica, por se tratar
de uma fonte renovável de energia, com altíssimo potencial produtivo e aliado a relativa
carência de pesquisa sobre a aplicação do sensoriamento remoto na optimização da cadeia
produtiva desta cultura, fica evidente a necessidade de investigações neste sentido, que possam
vir a oferecer alternativas tecnológicas no auxilio do monitoramento desta cultura.
Neste sentido o presente trabalho visou desenvolver um modelo a partir de um sensor
hiperespectral de campo para a estimativa do Teor Relativo de Clorofila (TRC) tendo como
base, dados mensurados por um clorofilômetro portátil.
2. Metodologia de Trabalho
O estudo foi desenvolvido em um canavial comercial localizado no município de
Pradópolis-SP, nas coordenadas 21º15’08” S e 48º06’56” O. E experimento foi instalado em
um Latossolo Vermelho, cultivado com a variedade CTC2, em quarto corte. O delineamento
experimental foi em blocos ao acaso, com quatro repetições. Os tartamentos consisitiram de
doses de N (0, 60, 120 e 180 kg N ha-1), aplicados em parcelas com dimensões de 10 m por 5
linhas de plantio. A fonte N foi o Nitrato de Amônio.
Para aquisição dos dados espectrais foi utilizado o aparelho FieldSpec 4 spectroradiometer,
com resolução espectral de 1 nm de 400 a 2500 nm ( ). O aparelho foi posicionado um metro
acima do dossel, quando as plantas da área experimental possuiam em média 0,7 metro de altura
de colmos, em cinco pontos previamente demarcados dentro da área útil da parcela (Figura 1).
Todas as avaliações foram realizadas no dia 19/12/2012.
Devido a interação das moléculas de água presente na atmosfera com a energia
eletromagnética, as curvas espectrais obtidas a campo, apresentam alto nível de ruído nos
comprimentos de onda centrados em 1400 nm e 1900 nm, portanto estas regiões foram
aliminadas do conjunto de dados.
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6314
Figura 1. Esquema de uma subparcela da área experimental e plano de avaliação.
As medidas do TRC, foram realizadas utilizando o clorofilômetro SPAD-502 (Spectrum
Technologies, Plainfield, IL, USA), em quatro folhas diagnóstico “+1”, para cada ponto
amostrado pelo espectrorradiometro. Tanto para os dados espectrais como para o TRC foram
utilizados os valores médios de cada ponto para análise. O TRC não possui unidade definida,
mas valores elevados estão diretamente relacionados com o teor absoluto de clorofila.
A partir das curvas espectrais, iniciou-se a seleção das bandas que melhor explicam as
variações do TRC nos dados analisados. Primeiramente realizou-se a análise de correlação de
cada comprimento de onda com os resultados do TRC. Para testar a significância da correlação,
aplicou-se o teste t de student, sendo que nesta fase foram mantidos apenas os comprimentos
de onda que apresentaram significância igual ou maior ao nível de 5% de probabilidade.
Posteriormente, partindo dos comprimentos de onda mais correlacionados com o TRC, foi
realizada a correlação com as demais bandas, sendo que quando estas mostravam coeficientes
de correlação maiores que 0,8, o comprimento de onda com menor correlação com a variável
predita era eliminada. Este processo foi realizado de modo a selecionar as bandas com maiores
correlações significativas com a variável predita e ao mesmo tempo eliminar as variáveis
preditoras com alta intercorrelação, reduzindo assim os efeitos da multicolinearidade.
Utilizando dados da região do vermelho e do infravermelho próximo, foi calculado também
o Índice de Vegetação de Diferença Normalizada (NDVI), conforme proposto por Lichtenthaler
et al. (1996), calculado pela equação 1.
NDVI = 750 nm550 nm
750 nm + 550 nm
Com as variáveis selecionadas foi gerado um modelo de regressão linear multipla por
stepwise para predição do teor relativo de clorofila, onde dos 80 pontos amostrados na área
experimental, foram utilizados dados de 64 pontos para geração do modelo e os 16 restantes
foram separados aleatoriamente do conjunto inicial e utilizados posteriormente na fase de
validação do modelo. A eficiência do modelo de validação, foi avaliado de acordo com o
coeficiente de determinação, erro médio de predição e a proximidade da linha de tendência
central da disperção entre dados estimados e observados, em relação a linha 1:1.
3. Resultados e Discussão
10 m
1,5 m
1 m
6 m
Ponto demarcado
Leituras FieldSpec
Área avaliada
Área útil
1
2
3
4 5
(1)
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6315
O aumento das doses de nitrogênio ocasionou redução na reflectância das folhas (Figura 2,
com redução mais acentuada na região do verde e do infravermelho próximo. Apesar da
clorofila absorver menos energia eletromagnética na região do azul e vermelho, o aumento da
concentração deste pigmento reduz a energia refletida em toda a região do visível (LAMB et
al., 2002), o tratamento que não recebeu aplicação de nitrogênio, mostrou maiores reflectância
nesta região, mais evidente no verde e na transição entre verde e vermelho (520 a 650 nm), o
que visualmente ocasiona uma coloração verde clara a amarelada uniforme, típica da
deficiência de nitrogênio.
Figura 2. Curva espectral média do dossel da cultura da cana-de-açúcar, sob aplicação de
diferentes doses de nitrogênio.
Para a região do infravermelho próximo as doses também tiveram ação inversa sobre a
reflectância do dossel, ou seja, maiores doses ocasionaram menor reflectância nesta região
espectral. Esta região é afetada principalmente pela estrutura celular da vegetação, diferença de
refração entre o conteúdo de água e ar presente na folha, e a reflectância aditiva, diretamente
ligada ao Índice de Área Foliar (DATT, 1999).
Para a faixa do espectro analisada, foi observada correlação significativa no intervalo entre
520-610 nm e 720 a 1070 nm, tanto para com as doses de nitrogênio como para o TRC. Nas
demais regiões as correlações foram fracas, apresentando além disso, muitos ruídos aleatórios
e, portanto, estes dados foram excluídos das análises posteriores (Figura 3). A maior parte dos
modelos e índices disponíveis na literatura, apontam comprimentos de ondas similares como os
mais significativos para a predição do nitrogênio na vegetação (RANJAN et al., 2012).
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6316
Figura 3. Correlação das doses de N aplicadas e teor relativo de clorofila medida pelo
clorofilômetro SPAD-502 com cada banda do sensor FiledSpec.
Os comprimentos de onda mais significativos para a avaliação do TRC foram o 550 nm
(verde), a região de transição entre o visível e o infravermelho próximo, conhecida como red-
edge (725 nm) (Boochs et al., 1990) e outra na região do infravermelho próximo (750 nm), com
coeficientes de correlação de -0,50, -0,69 e -0,77 respectivamente (Tabela 1).
Os resultados obtidos pelo cálculo do NDVI, mostraram correlação significativa de 0.68
com o TRC (Tabela 1). Embora os resultados sejam inferiores aos observados para os valores
de reflectância absoluta em diferentes comprimentos de onda, os índices tendem a ser mais
estáveis quando avaliações em diferentes espécies, culturas, variedades e ambientes de cultivos,
uma vez que a relação entre bandas são menos sensíveis a tais variações (LI et al., 2010)
Tabela 1. Matriz de correlação entre o Teor Relativo de Clorofila (TRC) com as bandas
selecionadas e o Índice de Vegetação de Diferença Normalizada (NDVI).
A combinação linear das variáveis selecionadas por regressão linear múltipla por stepwise
mostrou um coeficiente de determinação (R²) de 0,69 com os valores do TRC medidos pelos
clorofilômetro SPAD-502, com erro médio de 0.81 TRC (valor do SPAD-502) (Tabela 2). O
modelo combinado entre valores de reflectância das bandas e o NDVI, foi estatisticamente
melhor que os modelos utilizando apenas as bandas ou o índice separadamente, por esta razão
apenas os resultados para o modelo combinado será apresentado.
Correlação
significativa
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6317
Tabela 2. Parâmetros descritivos da equação para determinação do teor relativo de clorofila por
dados espectrais de dossel com as bandas selecionadas (550, 725 e 750 nm) e o NDVI.
A validação do modelo mostrou que a correlação entre os valores estimados pelo modelo
espectral de dossel e os valores do TRC medidos pelo aparelho SPAD-502 foi eficiente,
apresentando coeficiente de correlação significativo de 0,8, com a linha de tendência entre os
valores estimados e observados muito próximos a linha 1:1 como exibido no gráfico de
dispersão (Figura 4). Este comportamento foi observado por outros autores, como
apresentado por Miao et al. (2008), que porém obteve melhores resultados na fase de calibração
e validação, estudando a cultura do milho em duas áreas no estado de Minnesota nos Estados
Unidos.
Figura 4. Validação do modelo espectral para predição do teor relativo de clorofila.
Apesar do pequeno universo amostral do presente trabalho, observa-se que a resposta
espectral do dossel da cana-de-açúcar é sensível a variação de pigmentos da folha que por sua
vez está diretamente ligado ao estado nutricional de nitrogênio na vegetação. Deste modo o
conhecimento detalhado desta relação poderá auxiliar de forma significativa o monitoramento
e gerenciamento da aplicação deste nutriente. Este conhecimento fornece subsídios para a
realização de trabalhos mais robustos, que englobem mais variedades, ambientes de produção
e condições mais extremas de estresse de nitrogênio na cultura.
R-Quadrado 0,77
R-quadrado ajustado 0,69
Erro padrão 0,81
Observações 64
Modelo
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6318
4. Conclusões
Os resultados obtidos neste trabalho sugerem que existe boa correlação entre a resposta
espectral do dossel e o teor de clorofila presente na cultura da cana-de-açúcar, sendo as regiões
espectrais mais importantes o verde (550 nm), red-edge (720 nm) e infravermelho próximo
(750 nm).
Agradecimentos
À agência de Financiamento de Estudos e Projetos (FINEP), pelo financiamento do projeto ao
qual pertence a presente pesquisa.
Referências Bibliográficas
Boochs, F., Kupfer, G., Dockter, K., Kühbauch, W. Shape of the red-edge as vitality indicator for plants.
International Journal of Remote Sensing, v. 11, p. 17411753, 1990.
Chen, P., Haboudane, D., Tremblay, N., Wang, J., Vigneault, P., Li, B. New spectral indicator assessing the
efficiency of crop nitrogen treatment in corn and wheat. Remote Sensing of Environment, v. 114, n. 9, p.
19871997, 2010.
Datt, B. Visible/near infrared reflectance and chlorophyll content in Eucalyptus leaves. International Journal of
Remote Sensing, v. 20, n. 14, p. 27412759, 1999.
Lamb, D. W., Steyn-Ross, M., Schaare, P., Hanna, M. M. Estimating leaf nitrogen concentration in ryegrass
(Lolium spp.) pasture using the chlorophyll rededge Theoretical modeling and experimental observations.
International Journal of Remote Sensing, v. 23, p. 36193648, 2002.
Li, F., Miao, Y., Hennig, S. D., Gnyp, M. L., Chen, X., Jia, L., Bareth, G. Evaluating hyperspectral vegetation
indices for estimating nitrogen concentration of winter wheat at different growth stages. Precision Agriculture,
v. 11, n. 4, p. 335357, 2010.
Lichtenthaler, H. K., Lang, M., Sowinska, M., Heisel, F., Mieh, J. A. Detection of vegetation stress via a new
high-resolution fluorescence imaging system. Journal of Plant Physiology, v. 148, p. 599612, 1996.
Miao, Y., Mulla, D. J., Randall, G. W., Vetsch, J., Vintila, R. Combining chlorophyll meter readings and high
spatial resolution remote sensing images for in-season site-specific nitrogen management of corn. Precision
Agriculture, v. 10, n. 1, p. 4562, 2008.
Miphokasap, P., Honda, K., Vaiphasa, C., Souris, M., Nagai, M. Estimating Canopy Nitrogen Concentration in
Sugarcane Using Field Imaging Spectroscopy. Remote Sensing, v. 4, n. 12, 16511670, 2012.
Portz, G., Molin, J. P., Jasper, J. Active crop sensor to detect variability of nitrogen supply and biomass on
sugarcane fields. Precision Agriculture, v. 13, n. 1, p. 3344.
Ranjan, R., Chopra, U. K., Sahoo, R. N., Singh, A. K., Pradhan, S. Assessment of plant nitrogen stress in wheat (
Triticum aestivum L.) through hyperspectral indices. International Journal of Remote Sensing, v. 33, n. 20,
63426360, 2012.
Song, S., Gong, W., Zhu, B., Huang, X. Wavelength selection and spectral discrimination for paddy rice, with
laboratory measurements of hyperspectral leaf reflectance. ISPRS Journal of Photogrammetry and Remote
Sensing, v. 66, n. 5, p. 672682 2011.
Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE
6319
... Recently, proximal sensing has been gaining adherents in Brazil, especially in sugarcane fields [22,23], and later in grain fields, such as soybean [24]. Although precision tools currently are used more for planting/sowing operations, fertilization and weed control, there is a growing interest of researchers in providing tools to be used in insect and plant pathogen (disease) management [25]. ...
Article
Full-text available
Although monitoring insect pest populations in the fields is essential in crop management, it is still a laborious and sometimes ineffective process. Imprecise decision-making in an integrated pest management program may lead to ineffective control in infested areas or the excessive use of insecticides. In addition, high infestation levels may diminish the photosynthetic activity of soybean, reducing their development and yield. Therefore, we proposed that levels of infested soybean areas could be identified and classified in a field using hyperspectral proximal sensing. Thus, the goals of this study were to investigate and discriminate the reflectance characteristics of soybean non-infested and infested with Bemisia tabaci using hyperspectral sensing data. Therefore, cages were placed over soybean plants in a commercial field and artificial whitefly infestations were created. Later, samples of infested and non-infested soybean leaves were collected and transported to the laboratory to obtain the hyperspectral curves. The results allowed us to discriminate the different levels of infestation and to separate healthy from whitefly infested soybean leaves based on their reflectance. In conclusion, these results show that hyperspectral sensing can potentially be used to monitor whitefly populations in soybean fields.
... Recently, geotechnology has been gaining adherents in Brazil, especially in sugarcane fields [30,31], and later in grain fields, such as soybean [32]. Although precision tools currently are more used for planting/sowing operations, fertilization and weed control, there is a growing interest of researchers in providing tools to be used in insect and plant pathogen (disease) management [33]. ...
Preprint
Full-text available
Although monitoring and observing insect pest populations in the fields is essential in crop management, it is still a laborious and sometimes ineffective process. High infestation levels may diminish the photosynthetic activity of soybean plants, affecting their development and reducing the yield. An imprecise decision making in integrated pest management program may lead to an ineffective control in infested areas or the excessive use of insecticides. In order to reach a more efficient control of arthropods population it is important to evaluate the infestation in time to mitigate its negative effects on the crop and remote sensing is an important tool for monitoring. It was proposed that infested soybean areas could be identified, and the arthropods quantified from non-infested areas in a field by hyperspectral remote sensing. Thus, the goals of this study were to investigate and discriminate the reflectance characteristics of soybean non-infested and infested with Bemisia tabaci using hyperspectral remote sensing data. Therefore, samples of infested and non-infested soybean leaves were collected and transported to the laboratory to obtain the hyperspectral curves. The results obtained allowed to discriminate the different levels of infestation and to separate healthy from whitefly infested soybean leaves based on their reflectance.
... Nesse contexto, uma maior absorção representa a presença de pigmentos na folha, dentre eles clorofila, carotenóides e antocianina, e consequente maior taxa fotossintética (PONZONI et al., 2012). Pode-se também observar uma maior absorção de energia da plântula transgênica centrada em 550 nm (verde), referente a maiores teores de clorofila (FIORIO et al., 2015). Outra característica que justifica tal diferença nas curvas da plântula transgênica sob a convencional é maturidade da folha, devido à presença de espaços no mesófilo que absorvem mais luz nessa faixa (LIU, 2006), resultado da alta velocidade de germinação. ...
Conference Paper
Full-text available
In the recent Brazilian agricultural scenario, soybean crop has been high relevance, especially in its transgenic variable, which reaches high levels of productivity. With these genetic modifications, there is a need a new methodology of fast and non-destructive way to differentiate conventional and modified soybeans, that is the spectroscopy. The objective of this study was to analyze soybean seedlings spectral response to identify if there are differences between conventional and transgenic variables. Conventional and transgenic soybean seedlings were cultivated in greenhouses at ESALQ/USP, collected 15 days after emergence. Seedlings radiometric response were performed through Fieldspec 3 Jr. On the basis of these data, two mean spectral curves were elaborated in the EXCEL software. The results showed differences in the VIS and NIR ranges, due to transgenic plants physiology.
Article
Full-text available
A field experiment with wheat was conducted with four different nitrogen and four different water stress levels, and hyperspectral reflectances in the 350–2500 nm range were recorded at six crop phenostages for two years (2009–2010 and 2010–2011). Thirty-two hyperspectral indices were determined using the first-year reflectance data. Plant nitrogen (N) status, characterized by leaf nitrogen content (LNC) and plant nitrogen accumulation (PNA), showed the highest R with the spectral indices at the booting stage. The best five predictive equations for LNC were based on the green normalized difference vegetation index (GNDVI), normalized difference chlorophyll index (NDCI), normalized difference705 (ND705) index, ratio index-1dB (RI-1dB) and Vogelman index a (VOGa). Their validation using the second-year data showed high R (>0.80) and ratio of performance to deviation (RPD; >2.25) and low root mean square error (RMSE;
Article
Full-text available
The chlorophyll meter (CM) has been commonly used for in-season nitrogen(N) management of corn (Zea mays L.). Nevertheless, it has limited potential for site specific N management in large fields due to difficulties in using it to generate N status maps. The objective of this study was to determine how well CM readings can be estimated using aerial hyperspectral and simulated multispectral remote sensing images at different corn growth stages. Two field experiments were conducted in Minnesota, USA during 2005 involving different N application rates and timings on a corn-soybean [Glycine max(L.) Merr.] rotation field and a corn-corn rotation field. Four flights were made during the growing season using the AISA Eagle Hyper-spectral Imager and CM readings were collected at four or five different growth stages. The results indicated that single multispectral and hyperspectral band or vegetation index could explain 64–86% and 73–88% of the variability in CM readings, respectively, except at growth stage V9 in the corn-soybean rotation field where no band or vegetation index could explain more than 37% of the variability in CM readings. Multiple regression analysis demonstrated that the combination of 2–4 broad-bands or 3–8 narrow-bands could explain 41–92% or 61–94% of the variability in CM readings across the two fields and different corn growth stages investigated. It was concluded that the combination of CM readings with high spatial resolution hyperspectral or multispectral remote sensing images can overcome the limitations of using them individually, thus offering a practical solution to N deficiency detection and possibly in-season site-specific N management in large continuous corn fields or at later stages in corn-soybean rotation fields.
Article
Full-text available
The retrieval of nutrient concentration in sugarcane through hyperspectral remote sensing is widely known to be affected by canopy architecture. The goal of this research was to develop an estimation model that could explain the nitrogen variations in sugarcane with combined cultivars. Reflectance spectra were measured over the sugarcane canopy using a field spectroradiometer. The models were calibrated by a vegetation index and multiple linear regression. The original reflectance was transformed into a First-Derivative Spectrum (FDS) and two absorption features. The results indicated that the sensitive spectral wavelengths for quantifying nitrogen content existed mainly in the visible, red edge and far near-infrared regions of the electromagnetic spectrum. Normalized Differential Index (NDI) based on FDS(750/700) and Ratio Spectral Index (RVI) based on FDS(724/700) are best suited for characterizing the nitrogen concentration. The modified estimation model, generated by the Stepwise Multiple Linear Regression (SMLR) technique from FDS centered at 410, 426, 720, 754, and 1,216 nm, yielded the highest correlation coefficient value of 0.86 and Root Mean Square Error of the Estimate (RMSE) value of 0.033%N (n = 90) with nitrogen concentration in sugarcane. The results of this research demonstrated that the estimation model developed by SMLR yielded a higher correlation coefficient with nitrogen content than the model computed by narrow vegetation indices. The strong correlation between measured and estimated nitrogen concentration indicated that the methods proposed in this study could be used for the reliable diagnosis of nitrogen quantity in sugarcane. Finally, the success of the field spectroscopy used for estimating the nutrient quality of sugarcane allowed an additional experiment using the polar orbiting hyperspectral data for the timely determination of crop nutrient status in rangelands without any requirement of prior cultivar information.
Article
Full-text available
Chlorophyll red-edge descriptors have been used to estimate leaf nitro-gen concentration in ryegrass (L olium spp.) pasture. Two-layer model calculations have been used to predict the inn uence of chlorophyll content and Leaf Area Index (LAI) on the shape and location of the peaks observed in the derivative spectra of a ryegrass canopy. The complex structure of the resulting derivative spectra pre-cluded extracting red-edge wavelengths by tting inverted Gaussian curves to ree ectance proo les. Fitting a combination of three sigmoid curves to the calculated ree ectance spectra provided a better representation of subsequent derivative spectra. The derivative spectra in the vicinity of the chlorophyll red-edge is predicted to contain two peaks (~705 and ~725 nm), which on increasing the canopy LAI is generally found to shift to longer wavelengths. However, for a canopy containing leaves of low chlorophyll content and LAI>5, the wavelength of the rst peak becomes insensitive to changes in LAI. The same phenomenon is predicted for high-chlorophyll leaves of LAI>10. The role of multiple scattering, primarily due to increased leaf transmittance at higher wavelengths, has also been verii ed. In subsequent experiments, the predicted shape of the derivative spectra was observed and the use of three sigmoid curves to better represent this shape veri ed. Changes in the descriptors used to describe the chlorophyll red-edge were observed to explain 60% and 65% of the variance of leaf nitrogen concentration and total leaf nitrogen content, respectively. The resulting regression equation was found to predict leaf nitrogen concentration, in the range of 2–5.5%, with a standard error of prediction (SEP) of 0.4%. The confounding inn uence of canopy biomass on the red-edge determination of leaf nitrogen concentration was found to be signii cantly less at higher canopy biomass, conn rming both theoretical predictions and the potential of using the chlorophyll red-edge as a biomass-independent means of estimating leaf chlorophyll, and hence nitrogen, concentration in high-LAI ryegrass pastures.
Article
Full-text available
Many hyperspectral vegetation indices (VIs) have been developed to estimate crop nitrogen (N) status at leaf and canopy levels. However, most of these indices have not been evaluated for estimating plant N concentration (PNC) of winter wheat (Triticum aestivum L.) at different growth stages using a common on-farm dataset. The objective of this study was to evaluate published VIs for estimating PNC of winter wheat in the North China Plain for different growth stages and years using data from both N experiments and farmers’ fields, and to identify alternative promising hyperspectral VIs through a thorough evaluation of all possible two band combinations in the range of 350–1075 nm. Three field experiments involving different winter wheat cultivars and 4–6 N rates were conducted with cooperative farmers from 2005 to 2007 in Shandong Province, China. Data from 69 farmers’ fields were also collected to evaluate further the published and newly identified hyperspectral VIs. The results indicated that best performing published and newly identified VIs could explain 51% (R700/R670) and 57% (R418/R405), respectively, of the variation in PNC at later growth stages (Feekes 8–10), but only 22% (modified chlorophyll absorption ratio index, MCARI) and 43% (R763/R761), respectively, at the early stages (Feekes 4–7). Red edge and near infrared (NIR) bands were more effective for PNC estimation at Feekes 4–7, but visible bands, especially ultraviolet, violet and blue bands, were more sensitive at Feekes 8–10. Across site-years, cultivars and growth stages, the combination of R370 and R400 as either simple ratio or a normalized difference index performed most consistently in both experimental (R 2 = 0.58) and farmers’ fields (R 2 = 0.51). We conclude that growth stage has a significant influence on the performance of different vegetation indices and the selection of sensitive wavelengths for PNC estimation, and new approaches need to be developed for monitoring N status at early growth stages.
Article
The UV-laser ( λ 355 nm) induced fluorescence emission spectra of green leaves comprise the blue(F440) and green (F520) fluorescence bands as well as the red (F690) and far-red (F740) chlorophyll fluorescence emission bands. Based on the four UV-laser induced fluorescence bands blue, green, red and farred a high resolution fluorescence imaging system was established, which allows a fast and large scale screening of fluorescence gradients and local disturbances in fluorescence emission over the whole leaf surface. The new imaging method not only permits to screen leaves by means of images in four fluorescence bands (LIF images) but, in addition, via images of the fluorescence ratios blue/red (F440/F690), blue/farred (F440/F740), the chlorophyll fluorescence ratio red/far-red (F690/F740) and the ratio blue/green (F440/F520) (LIF ratio images). By fluorescence imaging we could prove that in aurea tobacco the major part of the leaves' blue and green fluorescence is emitted from the main and side leave veins, whereas the major part of the leaves' red and far-red chlorophyll fluorescence is emitted from the vein-free leaf regions, which also have the highest chlorophyll content. A smaller proportion of the aurea tobacco leaves' blue-green fluorescence emission is derived from the cell walls of epidermis cells. The fluorescence ratios blue/red and blue/far-red are very sensitive to environmental changes, and thus permit early stress and strain detection in plants, and the evaluation of damage to the photosynthetic apparatus. Via monitoring the increase in chlorophyll fluorescence LIF images allow to detect differences in the time-dependent uptake of diuron and the progressing inhibition of photosynthetic electron transport in the treated leaf part. The novel fluorescence imaging technique sets a new dimension for early stress detection in the photosynthetic apparatus and in plants. It has many advantages over the previously applied point-data measurements of selected leaf points using conventional spectrofluorometers. The new fluorescence imaging system proved to be very suitable for remote sensing of plants in the near distance, and can be further developed for far distance remote sensing of the state of health of terrestrial vegetation. Some examples in the many possible ways of computer-aided fluorescence data processing (formation of different fluorescence ratios, screening of fluorescence profiles, histogramme plotting) are indicated.
Article
The objective of this research is to select the most sensitive wavelengths for the discrimination of the imperceptible spectral variations of paddy rice under different cultivation conditions. The paddy rice was cultivated under four different nitrogen cultivation levels and three water irrigation levels. There are 2151 hyperspectral wavelengths available, both in hyperspectral reflectance and energy space transformed spectral data. Based on these two data sets, the principal component analysis (PCA) and band-band correlation methods were used to select significant wavelengths with no reference to leaf biochemical properties, while the partial least squares (PLS) method assessed the contribution of each narrow band to leaf biochemical content associated with each loading weight across the nitrogen and water stresses. Moreover, several significant narrow bands and other broad bands were selected to establish eight kinds of wavelength (broad-band) combinations, focusing on comparing the performance of the narrow-band combinations instead of broad-band combinations for rice supervising applications. Finally, to investigate the capability of the selected wavelengths to diagnose the stress conditions across the different cultivation levels, four selected narrow bands (552, 675, 705 and 776nm) were calculated and compared between nitrogen-stressed and non-stressed rice leaves using linear discriminant analysis (LDA). Also, wavelengths of 1158, 1378 and 1965nm were identified as the most useful bands to diagnose the stress condition across three irrigation levels. Results indicated that good discrimination was achieved. Overall, the narrow bands based on hyperspectral reflectance data appear to have great potential for discriminating rice of differing cultivation conditions and for detecting stress in rice vegetation; these selected wavelengths also have great potential use for the designing of future sensors.
Article
The shift of the red edge in the reflection spectra of vegetation targets is a known phenomenon documenting changes in the biological status of plants. In our study we analysed the variability of red edge reflection in dependency of differently managed field plots, The results indicate that the red edge is not fully described by the shift of the main inflection point, but has to be considered as a collection of several different and possibly independent features, each of them influenced by biological parameters of the plants. Thus, taking all features, the red edge as derived from high resolution spectra may provide enough information to detect small differences in the chemical and morphological status of plants.
Article
Nitrogen management has been intensively studied on several crops and recently associated with variable rate on-the-go application based on crop sensors. Such studies are scarce for sugarcane and as a biofuel crop the energy input matters, seeking high positive energy balance production and low carbon emission on the whole production system. This article presents the procedure and shows the first results obtained using a nitrogen and biomass sensor (N-Sensor™ ALS, Yara International ASA) to indicate the nitrogen application demands of commercial sugarcane fields. Eight commercial fields from one sugar mill in the state of São Paulo, Brazil, varying from 15 to 25 ha in size, were monitored. Conditions varied from sandy to heavy soils and the previous harvesting occurred in May and October 2009, including first, second, and third ratoon stages. Each field was scanned with the sensor three times during the season (at 0.2, 0.4, and 0.6 m stem height), followed by tissue sampling for biomass and nitrogen uptake at ten spots inside the area, guided by the different values shown by the sensor. The results showed a high correlation between sensor values and sugarcane biomass and nitrogen uptake, thereby supporting the potential use of this technology to develop algorithms to manage variable rate application of nitrogen for sugarcane.
Article
The visible/near infrared reflectance properties of leaves from several Eucalyptus species were studied to determine appropriate indices for remote sensing of chlorophyll content. A scatter correction technique was applied to the reflectance spectra to reduce the additive and multiplicative scattering effects of leaf surface and internal structure. This method gave improved calibrations for the estimation of chlorophyll content. Reflectance near 710nm wavelength showed maximum sensitivity to chlorophyll content. Reflectance near 550nm was a less sensitive indicator of chlorophyll content. Among several reflectance indices tested, the ratio (R850-R710)/(R850-R680) performed best, and is proposed as a new index for the remote estimation of chlorophyll content in higher plants. From the first derivative of reflectance, the ratio D1(754)/D1(704) and the wavelength position of the red edge were best correlated to chlorophyll content. The ratio D2(712)/D2(688) from the second derivative of reflectance was also an equally good indicator of chlorophyll content.