No abstract

Over use of nitrogen fertilization can result in groundwater pollution. Tools that can rapidly quantify the nitrogen status are needed for efficient fertilizer management and would be very helpful in reducing the environmental pollution caused by excessive nitrogen application. Remote sensing has a proven ability to provide spatial and temporal measurements of surface properties. In this study, the MLR (multiple linear regression) and ANN (artificial neural network) modeling methods were applied to the monitoring of rice N (nitrogen concentration, mg nitrogen g(-1) leaf dry weight) status using leaf level hyperspectral reflectance with two different input variables, and as a result four estimation models were proposed. RMSE (root-mean-square error), REP (relative error of prediction), R2 (coefficient of determination), as well as the intercept and slope between the observed and predicted N were used to test the performance of models. Very good agreements between the observed and the predicted N were obtained with all proposed models, which was especially true for the R-ANN (artificial neural network based on reflectance selected using MLR) model. Compared to the other three models, the R-ANN model improved the results by lowering the RMSE by 14.2%, 32.1%, and 31.5% for the R-LR (linear regression based on reflectance) model, PC-LR (linear regression based on principal components scores) model, and PC-ANN (artificial neural network based on principal components scores) model, respectively. It was concluded that the ANN algorithm may provide a useful exploratory and predictive tool when applied on hyperspectral reflectance data for nitrogen status monitoring. Besides, although the performance of MLR was superior to PCA used for ANN inputs selection, the encouraging results of PC-based models indicated the promising potential of ANN combined with PCA application on hyperspectral reflectance analysis.

No abstract

There has been a great deal of interests in the estimation of grassland biophysical parameters such as percentage of vegetation cover (PVC), aboveground biomass, and leaf-area index with remote sensing data at the canopy scale. In this paper, the percentage of vegetation cover was estimated from vegetation indices using Moderate Resolution Imaging Spectroradiometer (MODIS) data and red-edge parameters through the first derivative spectrum from in situ hypserspectral reflectance data. Hyperspectral reflectance measurements were made on grasslands in Inner Mongolia, China, using an Analytical Spectral Devices spectroradiometer. Vegetation indices such as the difference, simple ratio, normalized difference, renormalized difference, soil-adjusted and modified soil-adjusted vegetation indices (DVI, RVI, NDVI, RDVI, SAVI L = 0.5 and MSAVI 2 ) were calculated from the hyperspectral reflectance of various vegetation covers. The percentage of vegetation cover was estimated using an unsupervised spectral-contextual classifier automatically. Relationships between percentage of vegetation cover and various vegetation indices and red-edge parameters were compared using a linear and second-order polynomial regression. Our analysis indicated that MSAVI 2 and RVI yielded more accurate estimations for a wide range of vegetation cover than other vegetation indices and red-edge parameters for the linear and second-order polynomial regression, respectively. Key words: grassland; hypserspectral remote sensing; percentage of vegetation cover; red-edge parameter; vegetation index. Liu ZY, Huang JF, Wu XH, Dong YP (2007). Comparison of vegetation indices and red-edge parameters for estimating grassland cover from canopy reflectance data.Estimation of vegetation cover is very important for assessing livestock-carrying capacity of grasslands and monitoring the extent of desertification in arid and semi-arid areas. Green vegetation cover is also an important factor in soil erosion control (Purevdorj et al.1998;Fan et al. 2002), and, therefore, a sensitive indicator of land degradation and desertification. Furthermore, quantitative information on the green vegetation cover is required in global carbon cycle studies and climate change research at the global and local scale (Qi et al.1994;Boles et al. 2004).Generally, there are two types of approaches to quantifying percentages of vegetation cover: the traditional ground-based observation versus the remote sensing approach. The former includes visual estimation, instrumental measurements and destructive sampling, which have exerted a very important value in studies of impact of climate change and plant community ecology. However, the ground-based observations are costly and time-consuming, and are difficult to execute in remotely

Detection of crop health conditions plays an important role in making control strategies of crop disease and insect damage and gaining high-quality production at late growth stages. In this study, hyperspectral reflectance of rice panicles was measured at the visible and near-infrared regions. The panicles were divided into three groups according to health conditions: healthy panicles, empty panicles caused by Nilaparvata lugens Stål, and panicles infected with Ustilaginoidea virens. Low order derivative spectra, namely, the first and second orders, were obtained using different techniques. Principal component analysis (PCA) was performed to obtain the principal component spectra (PCS) of the foregoing derivative and raw spectra to reduce the reflectance spectral dimension. Support vector classification (SVC) was employed to discriminate the healthy, empty, and infected panicles, with the front three PCS as the independent variables. The overall accuracy and kappa coefficient were used to assess the classification accuracy of SVC. The overall accuracies of SVC with PCS derived from the raw, first, and second reflectance spectra for the testing dataset were 96.55%, 99.14%, and 96.55%, and the kappa coefficients were 94.81%, 98.71%, and 94.82%, respectively. Our results demonstrated that it is feasible to use visible and near-infrared spectroscopy to discriminate health conditions of rice panicles.

Abstract:Detecting plant health conditions plays a key role in farm pest management and crop protection. In this study, measurement of hyperspectral leaf reflectance in rice crop (Oryzasativa L.) was conducted on groups of healthy and infected leaves by the fungus Bipolaris oryzae (Helminthosporium oryzae Breda. de Hann) through the wavelength range from 350 to 2 500 nm. The percentage of leaf surface lesions was estimated and defined as the disease severity. Statistical methods like multiple stepwise regression, principal component analysis and partial least-square regression were utilized to calculate and estimate the disease severity of rice brown spot at the leaf level. Our results revealed that multiple stepwise linear regressions could efficiently estimate disease severity with three wavebands in seven steps. The root mean square errors (RMSEs) for training (n=210) and testing (n=53) dataset were 6.5% and 5.8%, respectively. Principal component analysis showed that the first principal component could explain approximately 80% of the variance of the original hyperspectral reflectance. The regression model with the first two principal components predicted a disease severity with RMSEs of 16.3% and 13.9% for the training and testing dataset, respectively. Partial least-square regression with seven extracted factors could most effectively predict disease severity compared with other statistical methods with RMSEs of 4.1% and 2.0% for the training and testing dataset, respectively. Our research demonstrates that it is feasible to estimate the disease severity of rice brown spot using hyperspectral reflectance data at the leaf level.

No abstract

Part 1: GIS, GPS, RS and Precision FarmingInternational audienceHyperspectral reflectance of normal and lodged rice caused by rice brown planthopper and rice panicle blast was measured at the canopy level. Over one decade broad- and narrow-band vegetation indices (VIs) were calculated to simulate Landsat ETM+ with in situ hyperspectral reflectance. Principal component analysis (PCA) was utilized to obtain the front two principal components (PCs). Probabilistic neural network (PNN) was employed to classify the lodged and normal rice with VIs and PCs as the input vectors. PCs had 100% of overall accuracy and 1 of Kappa coefficient for the training dataset. While PCs had the greatest average overall accuracy (97.8%) and Kappa coefficient (0.955) for the two testing datasets than VIs consisting of broad- and narrow-bands. The results indicated that hyperspectral remote sensing with PCA and artificial neural networks could potentially be applied to discriminate lodged crops from normal ones at regional and large spatial scales