In situ hyperspectral data analysis for pigment content estimation of rice leaves

ArticleinJournal of Zhejiang University SCIENCE 4(6):727-33 · November 2003with8 Reads
DOI: 10.1631/jzus.2003.0727 · Source: PubMed
Analyses of the correlation between hyperspectral reflectance and pigment content including chlorophyll-a, chlorophyll-b and carotenoid of leaves in different sites of rice were reported in this paper. The hyperspectral reflectance of late rice during the whole growing season was measured using a Spectroradiometer with spectral range of 350-1050 nm and resolution of 3 nm. The chlorophyll-a, chlorophyll-b and carotenoid contents in rice leaves in rice fields to which different levels of nitrogen were applied were measured. The chlorophyll-a content of upper leaves was well correlated with the spectral variables. However, the correlation between both chlorophyll-b and caroteniod and the spectral variables was far from that of chlorophyll-a. The potential of hyperspectral reflectance measurement for estimating chlorophyll-a of upper leaves was evaluated using univariate correlation and multivariate regression analysis methods with different types of predictors. This study showed that the most suitable estimated model of chlorophyll-a of upper leaves was obtained by using some hyperspectral variables such as SD(r), SD(b) and their integration.
    • e, multivariate analysis, multiple linear regression (MLR), principal components analysis (PCA) and partial least square regression (PLSR) was applied to estimate the biophysical, physiological or biochemical characteristics of rice plants. Nguyen and Lee (2006) established the variation of crop growth and nitrogen status within a field using PLSR. Chen et al., (2003) compared the precision of the model for chlorophyll content between vegetation indices and MLR. Yi et al., (2007) compared the precision of the estimation model for nitrogen concentration depending on several multivariate analyses. Ryu et al., (2005 Ryu et al., ( , 2007) also evaluated a nitrogen content prediction model using airborne
    [Show abstract] [Hide abstract] ABSTRACT: Airborne hyperspectral remote sensing was adapted to establish a general-purpose model for quantifying nitrogen content of rice plants at the heading stage using three years of data. There was a difference in dry mass and nitrogen concentration due to the difference in the accumulated daily radiation (ADR) and effective cumulative temperature (ECT). Because of these environmental differences, there was also a significant difference in nitrogen content among the three years. In the multiple linear regression (MLR) analysis, the accuracy (coefficient of determination: R2, root mean square of error: RMSE and relative error: RE) of two-year models was better than that of single-year models as shown by R2≥0.693, RMSE≤1.405gm−2 and RE≤9.136%. The accuracy of the three-year model was R2=0.893, RMSE=1.092gm−2 and RE=8.550% with eight variables. When each model was verified using the other data, the range of RE for two-year models was similar or increased compared with that for single-year models. In the partial least square regression (PLSR) model for the validation, the accuracy of two-year models was also better than that of single-year models as R2≥0.699, RMSE≤1.611gm−2 and RE≤13.36%. The accuracy of the three-year model was R2=0.837, RMSE=1.401gm−2 and RE=11.23% with four latent variables. When each model was verified, the range of RE for two-year models was similar or decreased compared with that for single-year models. The similarities and differences of loading weights for each latent variable depending on hyperspectral reflectance might have affected the regression coefficients and the accuracy of each prediction model. The accuracy of the single-year MLR models was better than that of the single-year PLSR models. However, accuracy of the multi-year PLSR models was better than that of the multi-year MLR models. Therefore, PLSR model might be more suitable than MLR model to predict the nitrogen contents at the heading stage using the hyperspectral reflectance because PLSR models have more sensitive than MLR models for the inhomogeneous results. Although there were differences in the environmental variables (ADR and ECT), it is possible to establish a general-purpose prediction model for nitrogen content at the heading stage using airborne hyperspectral remote sensing.
    Article · Jun 2011
    • ndex that reduces the effect of differences in leaf surface reflec- tance, Sima et al. (2002) were able to significantly improve the correlations with chlorophyll content. Their results demonstrate that spectral indices can be applied across species with widely varying leaf structure without the necessity for extensive calibration for each species. Cheng (2003) showed that the most suitable estimated model of chlorophyll a of upper leaves was obtained by using some hyper-spectral variables such as SDr, SDb and their integration. Anatoly AG (2003) reported that reciprocal reflectance (R ) –1 in the spectral range from 520 to 550 nm and 695 to 705 nm related closely to the total chlorophyll cont
    [Show abstract] [Hide abstract] ABSTRACT: Our objective in this study was to develop spectral absorption indices for prediction of leaf chlorophyll concentration based on blue/yellow/red/ edge absorption spectrum. Two field experiments were conducted to study the response of chlorophyll index based on leaf absorption spectra to chlorophyll concentration in rice. The ultimate, penultimate and third expanded leaves were sampled for chlorophyll measurements and the absorption spectra of the leaves on the main stem for three rice varieties at different growth stages to select the absorption wavelength position near zero and develop better algorithms for estimating chlorophyll concentration. Some indices called blue/yellow/red/ edge absorption spectra chlorophyll index (BEACI/ YEACI/ REACI) were calculated from elected absorption wavelength positions. For the 1 st experiment the correlation coefficients were similar between chlorophyll concentration and single leaf spectral absorption and between chlorophyll concentration and these indices. But the chlorophyll concentration had significant correlations (P<0.01) to these indices than single leaf spectral absorption in the 2 nd experiment. The liner regression models with single leaf spectral absorption y = -2.271A 480.188 + 5.574A 651.232 -2.899A 753.552 -0.269, y= -4.079A 480.188 -2.233A 753.552 + 5.892A 663.239 + 0.547 and y = 4.217A 651.232 -0.718A 753.552 -2.897A 663.239 -0.399 had higher power prediction total chlorophyll, chlororphyll a and chlorophyll b concentrations, respectively. Compared with BEACI and REACI, stepwise regression analysis showed that YEACI 630.610 , YEACI 570.169 and YEACI 651.232 were good predictive power for predicting chlorophyll total concentration, chlorophyll a concentration and chlororphyll b concentration respectively.
    Article · Aug 2009 · Ying yong sheng tai xue bao = The journal of applied ecology / Zhongguo sheng tai xue xue hui, Zhongguo ke xue yuan Shenyang ying yong sheng tai yan jiu suo zhu ban
    • Log(1/R) (also called pseudo absorbance) is often employed because it provides a curve comparable to an absorption curve, with peaks occurring at the corresponding absorption wavelengths. The relationships between agronomic parameters and some specific narrow spectral bands have promoted the development of various models to estimate the agronomic parameters at leaf to canopy scales (Cheng et al., 2003). However, most of those models were developed using the means of regression analysis based on an assumption of a linear relationship between dependent and independent variables.
    [Show abstract] [Hide abstract] ABSTRACT: AbstractThe radial basis function (RBF) emerged as a variant of artificial neural network. Generalized regression neural network (GRNN) is one type of RBF, and its principal advantages are that it can quickly learn and rapidly converge to the optimal regression surface with large number of data sets. Hyperspectral reflectance (350 to 2 500 nm) data were recorded at two different rice sites in two experiment fields with two cultivars, three nitrogen treatments and one plant density (45 plants m−2). Stepwise multivariable regression model (SMR) and RBF were used to compare their predictability for the leaf area index (LAI) and green leaf chlorophyll density (GLCD) of rice based on reflectance (R) and its three different transformations, the first derivative reflectance (D1), the second derivative reflectance (D2) and the log-transformed reflectance (LOG). GRNN based on D1 was the best model for the prediction of rice LAI and GLCD. The relationships between different transformations of reflectance and rice parameters could be further improved when RBF was employed. Owing to its strong capacity for nonlinear mapping and good robustness, GRNN could maximize the sensitivity to chlorophyll content using D1. It is concluded that RBF may provide a useful exploratory and predictive tool for the estimation of rice biophysical parameters.
    Full-text · Article · Apr 2009
    • The spatial and temporal variations in the field of these variables must be determined in order to match the crop requirements as closely as possible. Different remote sensing applications have proved to be a potential source of reflectance data for estimation of several canopy variables related to biophysical, physiological or biochemical characteristics (Ahlrichs and Bauer, 1983; Serrano et al., 2000; Thenkabail et al., 2000; Cheng et al., 2003; Tang, et al., 2004; Zhang, et al., 2006). Various Vegetation Indices (VIs) have been related to crop variables such as biomass, leaf area, plant cover, leaf gap fraction, nitrogen, and chlorophyll in cereals (Best and Harlan,(NDVI; Rouse et al., 1974 ) are the best-known indices .
    [Show abstract] [Hide abstract] ABSTRACT: Hyperspectral reflectance (350∼2500 nm) data were recorded at two different sites of rice in two experiment fields including two cultivars, and three levels of nitrogen (N) application. Twenty-five Vegetation Indices (VIs) were used to predict the rice agronomic parameters including Leaf Area Index (LAI, m2 green leaf/m2 soil) and Green Leaf Chlorophyll Density (GLCD, mg chlorophyll/m2 soil) by the traditional regression models and Radial Basis Function Neural Network (RBF). RBF emerged as a variant of Artificial Neural Networks (ANNs) in the late 1980’s. A large variety of training algorithms has been tested for training RBF networks. In this study, Original RBF (ORBF), Gradient Descent RBF (GDRBF), and Generalized Regression Neural Network (GRNN) were employed. Results showed that green waveband Normalized Difference Vegetation Index (NDVIgreen) and TCARI/OSAVI have the best prediction power for LAI by exponent model and ORBF respectively, and that TCARI/OSAVI has the best prediction power for GLCD by exponent model and GDRBF. The best performances of RBF are compared with the traditional models, showing that the relationship between VIs and agronomic variables are further improved when RBF is used. Compared with the best traditional models, ORBF using TCARI/OSAVI improves the prediction power for LAI by lowering the Root Mean Square Error (RMSE) for 0.1119, and GDRBF using TCARI/OSAVI improves the prediction power for GLCD by lowering the RMSE for 26.7853. It is concluded that RBF provides a useful exploratory and predictive tool when applied to the sensitive VIs.
    Full-text · Article · May 2007
  • [Show abstract] [Hide abstract] ABSTRACT: Sensing of an intracardiac electrogram is critical to pacemaker design and operation. Characterization of an intracardiac electrogram sensing electrode has long been based on the assumption that the electrode is very small and has no effect on the field it senses. In fact, any intracardiac sensing electrode has a significantly large surface area where electrical charges are induced and disturb the sensed potential field. To account for the effect of electrode size and position on electrogram sensing, a surface charge integral equation is formulated, based an the current continuity equation in homogeneous myocardial medium. A numerical method, such as the method of moments, is then applied to obtain the solution of the integral equation, meanwhile the potential sensed by the electrode. As an application of the proposed technique, several electrode configurations have been analyzed and their relative electrogram sensitivities are compared
    Article · Jan 1995 · Ying yong sheng tai xue bao = The journal of applied ecology / Zhongguo sheng tai xue xue hui, Zhongguo ke xue yuan Shenyang ying yong sheng tai yan jiu suo zhu ban
  • [Show abstract] [Hide abstract] ABSTRACT: The MODIS (moderate-resolution imaging spectroradiometer) surface reflectance product MOD09, which was strictly corrected by NASA, can be widely applied to monitor the change of land vegetation. In this paper, a quasi and synchronous experiment of MODIS sensor was performed, and the rice leaf area index (LAI) and chlorophyll content (Chltot) were measured. The relationships between the vegetation indices (VIs) derived from MOD09 and the rice LAI and Chltot were analyzed, and the estimation models were established. The VIs values derived from MOD09 were higher than those in the first three bands of MOD09, and EVI value was lower than NDVI value. In comparing with other Vis, EVI had a better relationship with LAI at different rice growth stages. MOD09-EVI was selected to construct the estimation model of rice LAI, and validated by other in situ sampling plot data to be more precise, suggesting MOD09-EVI was the best index to monitor rice LAI. There was a significant correlation between MOD09-Red band and Chltot at early and medium rice growth stages. The estimation model of rice canopy Chltot based on MOD09-Red band was established and validated, and no models were significantly efficient, except the estimation model of Chltot at booting stage.
    Article · Sep 2006
Show more