Biwu Chen

Evaluation of hyperspectral LiDAR for monitoring rice leaf nitrogen by comparison with multispectral LiDAR and passive spectrometer

Fast and nondestructive assessment of leaf nitrogen concentration (LNC) is critical for crop growth diagnosis and nitrogen management guidance. In the last decade, multispectral LiDAR (MSL) systems have promoted developments in the earth and ecological sciences with the additional spectral information. With more wavelengths than MSL, the hyperspectral LiDAR (HSL) system provides greater possibilities for remote sensing crop physiological conditions. This study compared the performance of ASD FieldSpec Pro FR, MSL, and HSL for estimating rice (Oryza sativa) LNC. Spectral reflectance and biochemical composition were determined in rice leaves of different cultivars (Yongyou 4949 and Yangliangyou 6) throughout two growing seasons (2014–2015). Results demonstrated that HSL provided the best indicator for predicting rice LNC, yielding a coefficient of determination (R2) of 0.74 and a root mean square error of 2.80 mg/g with a support vector machine, similar to the performance of ASD (R2 = 0.73). Estimation of rice LNC could be significantly improved with the finer spectral resolution of HSL compared with MSL (R2 = 0.56).

Estimating the leaf nitrogen content of paddy rice by using the combined reflectance and laser-induced fluorescence spectra

Paddy rice is one of the most important crops in China, and leaf nitrogen content (LNC) serves as a significant indictor for monitoring crop status. A reliable method is needed for precise and fast quantification of LNC. Laser-induced fluorescence (LIF) technology and reflectance spectra of crops are widely used to monitor leaf biochemical content. However, comparison between the fluorescence and reflectance spectra has been rarely investigated in the monitoring of LNC. In this study, the performance of the fluorescence and reflectance spectra for LNC estimation was discussed based on principal component analysis (PCA) and back-propagation neural network (BPNN). The combination of fluorescence and reflectance spectra was also proposed to monitor paddy rice LNC. The fluorescence and reflectance spectra exhibited a high degree of multi-collinearity. About 95.38%, and 97.76% of the total variance included in the spectra were efficiently extracted by using the first three PCs in PCA. The BPNN was implemented for LNC prediction based on new variables calculated using PCA. The experimental results demonstrated that the fluorescence spectra (R2 = 0.810, 0.804 for 2014 and 2015, respectively) are superior to the reflectance spectra (R2 = 0.721, 0.671 for 2014 and 2015, respectively) for estimating LNC based on the PCA-BPNN model. The proposed combination of fluorescence and reflectance spectra can greatly improve the accuracy of LNC estimation (R2 = 0.912, 0.890 for 2014 and 2015, respectively).

Multispectral LiDAR Point Cloud Classification: A Two-Step Approach

Target classification techniques using spectral imagery and light detection and ranging (LiDAR) are widely used in many disciplines. However, none of the existing methods can directly capture spectral and 3D spatial information simultaneously. Multispectral LiDAR was proposed to solve this problem as its data combines spectral and 3D spatial information. Point-based classification experiments have been conducted with the use of multispectral LiDAR; however, the low signal to noise ratio creates salt and pepper noise in the spectral-only classification, thus lowering overall classification accuracy. In our study, a two-step classification approach is proposed to eliminate this noise during target classification: routine classification based on spectral information using spectral reflectance or a vegetation index, followed by neighborhood spatial reclassification. In an experiment, a point cloud was first classified with a routine classifier using spectral information and then reclassified with the k-nearest neighbors (k-NN) algorithm using neighborhood spatial information. Next, a vegetation index (VI) was introduced for the classification of healthy and withered leaves. Experimental results show that our proposed two-step classification method is feasible if the first spectral classification accuracy is reasonable. After the reclassification based on the k-NN algorithm was combined with neighborhood spatial information, accuracies increased by 1.50–11.06%. Regarding identification of withered leaves, VI performed much better than raw spectral reflectance, with producer accuracy increasing from 23.272% to 70.507%.

Analyzing the performance of fluorescence parameters in the monitoring of leaf nitrogen content of paddy rice

Leaf nitrogen content (LNC) is a significant factor which can be utilized to monitor the status of paddy rice and it requires a reliable approach for fast and precise quantification. This investigation aims to quantitatively analyze the correlation between fluorescence parameters and LNC based on laser-induced fluorescence (LIF) technology. The fluorescence parameters exhibited a consistent positive linear correlation with LNC in different growing years (2014 and 2015) and different rice cultivars. The R2 of the models varied from 0.6978 to 0.9045. Support vector machine (SVM) was then utilized to verify the feasibility of the fluorescence parameters for monitoring LNC. Comparison of the fluorescence parameters indicated that F740 is the most sensitive (the R2 of linear regression analysis of the between predicted and measured values changed from 0.8475 to 0.9226, and REs ranged from 3.52% to 4.83%) to the changes in LNC among all fluorescence parameters. Experimental results demonstrated that fluorescence parameters based on LIF technology combined with SVM is a potential method for realizing real-time, non-destructive monitoring of paddy rice LNC, which can provide guidance for the decision-making of farmers in their N fertilization strategies.

Estimating Rice Leaf Nitrogen Concentration: Influence of Regression Algorithms Based on Passive and Active Leaf Reflectance

Nitrogen (N) is important for the growth of crops. Estimating leaf nitrogen concentration (LNC) accurately and nondestructively is important for precision agriculture, reduces environmental pollution, and helps model global carbon and N cycles. Leaf reflectance, especially in the visible and near-infrared regions, has been identified as a useful indicator of LNC. Except reflectance passively acquired by spectrometers, the newly developed multispectral LiDAR and hyperspectral LiDAR provide possibilities for measuring leaf spectra actively. The regression relationship between leaf reflectance spectra and rice (Oryza sativa) LNC relies greatly on the algorithm adopted. It would be preferable to find one algorithm that performs well with respect to passive and active leaf spectra. Thus, this study assesses the influence of six popular linear and nonlinear methods on rice LNC retrieval, namely, partial least-square regression, least squares boosting, bagging, random forest, back-propagation neural network (BPNN), and support vector regression of different types/kernels/parameter values. The R2, root mean square error and relative error in rice LNC estimation using these different methods were compared through the passive and active spectral measurements of rice leaves of different varieties at different locations and time (Yongyou 4949, Suizhou, 2014, Yangliangyou 6, Wuhan, 2015). Results demonstrate that BPNN provided generally satisfactory performance in estimating rice LNC using the three kinds of passive and active reflectance spectra.

Effect of fluorescence characteristics and different algorithms on the estimation of leaf nitrogen content based on laser-induced fluorescence lidar in paddy rice

Paddy rice is one of the most significant food sources and an important part of the ecosystem. Thus, accurate monitoring of paddy rice growth is highly necessary. Leaf nitrogen content (LNC) serves as a crucial indicator of growth status of paddy rice and determines the dose of nitrogen (N) fertilizer to be used. This study aims to compare the predictive ability of the fluorescence spectra excited by different excitation wavelengths (EWs) combined with traditional multivariate analysis algorithms, such as principal component analysis (PCA), back-propagation neural network (BPNN), and support vector machine (SVM), for estimating paddy rice LNC from the leaf level with three different fluorescence characteristics as input variables. Then, six estimation models were proposed. Compared with the five other models, PCA-BPNN was the most suitable model for the estimation of LNC by improving R2 and reducing RMSE and RE. For 355, 460 and 556 nm EWs, R2 was 0.89, 0.80 and 0.88, respectively. Experimental results demonstrated that the fluorescence spectra excited by 355 and 556 nm EWs were superior to those excited by 460 nm for the estimation of LNC with different models. BPNN algorithm combined with PCA may provide a helpful exploratory and predictive tool for fluorescence spectra excited by appropriate EW based on practical application requirements for monitoring the N status of crops.

Monitoring of Paddy Rice Varieties Based on the Combination of the Laser-Induced Fluorescence and Multivariate Analysis

Paddy rice is one of three major cereal crops in China, and the number of the paddy rice variety is increasing rapidly. The paddy rice variety is strongly related to crop yield and is also difficult to classify by using the naked eyes. A reliable approach is essential for accurately identifying different paddy rice varieties. Laser-induced fluorescence (LIF) technology has been widely utilized in many fields due to its particular advantages (rapid, non-intrusive, and sensitive). Thus, LIF combined with multivariate analysis that contained principal component analysis (PCA) and support vector machine (SVM) was proposed and was attempted to be utilized to identify different paddy rice varieties in this investigation. These fluorescence spectra displayed a high degree of multi-collinearity, and about 96.58% of the total variance contained in the laser-induced fluorescence spectra which were excited by a 532-nm excited wavelength can be explained by using the first three principle components. A SVM model with the help of threefold cross validation was used for paddy rice variety identification based on new variables calculated utilizing PCA. The numerical and experimental results displayed by using a confusion matrix and the classification accuracy can reach up to 91.36%. Thus, LIF technology combined with multivariate analysis can provide researchers with a faster and more effective tool for identifying different paddy rice varieties.

Potential of vegetation indices combined with laser-induced fluorescence parameters for monitoring leaf nitrogen content in paddy rice

Nitrogen (N) is important for the growth of crops. Leaf nitrogen content (LNC) serves as a crucial indicator of the growth status of crops and can help determine the dose of N fertilizer. Laser-induced fluorescence (LIF) technology and the reflectance spectra of crops are widely used to detect the biochemical content of leaves. Many vegetation indices (VIs) and fluorescence parameters have been developed to estimate LNC. However, the comparison among VIs and between fluorescence parameters and VIs has been rarely studied in the estimation of LNC. In this study, the performances of several published empirical VIs and fluorescence parameters for the estimation of paddy rice LNC were analyzed using the support vector machine (SVM) algorithm. Then, the optimal VIs (TVI, MTVI1, MTVI2, and MSAVI) and fluorescence parameters (F735/F460 and F685/F460), which were suitable for LNC monitoring in this study, were chosen. In addition, the combination of the VIs and fluorescence parameters was proposed as the input variables in the SVM model and used to estimate the LNC. Experimental results exhibited the promising potential of the LIF technology combined with reflectance for the accurate estimation of LNC, which provided guidance for monitoring the LNC.

Analyzing the Effect of Fluorescence Characteristics on Leaf Nitrogen Concentration Estimation

Leaf nitrogen concentration (LNC) is a significant indicator of crops growth status, which is related to crop yield and photosynthetic efficiency. Laser-induced fluorescence is a promising technology for LNC estimation and has been widely used in remote sensing. The accuracy of LNC monitoring relies greatly on the selection of fluorescence characteristics and the number of fluorescence characteristics. It would be useful to analyze the performance of fluorescence intensity and ratio characteristics at different wavelengths for LNC estimation. In this study, the fluorescence spectra of paddy rice excited by different excitation light wavelengths (355 nm, 460 nm, and 556 nm) were acquired. The performance of the fluorescence intensity and fluorescence ratio of each band were analyzed in detail based on back-propagation neural network (BPNN) for LNC estimation. At 355 nm and 460 nm excitation wavelengths, the fluorescence characteristics related to LNC were mainly located in the far-red region, and at 556 nm excitation wavelength, the red region being an optimal band. Additionally, the effect of the number of fluorescence characteristics on the accuracy of LNC estimation was analyzed by using principal component analysis combined with BPNN. Results demonstrate that at least two fluorescence spectral features should be selected in the red and far-red regions to estimate LNC and efficiently improve the accuracy of LNC estimation.

Combined application of 3D spectral features from multispectral LiDAR for classification

Combining the multispectral rasterized data and the three-dimensional (3D) lidar point cloud has long been a hot topic in the remote sensing field. This facilitates not only target recognition, land-classification, but also understanding for the ecosystems and environment. To address this problem, the concept of novel multispectral lidar (MSL), which captures multispectral reflectance and accurate spatial traits simultaneously, was proposed in this study. The layout of the instrument was described. Four laser diodes were co-aligned into a single beam. The reflectance spectrum at four wavelengths (covering red-edge region) as well as distance were recorded. In a validation experiment, reflectance at four wavelengths and normal vectors obtained by the MSL system were fully utilized to classify different targets including fresh and sere plants, with an overall accuracy of 85.5%. The novel MSL was demonstrated to have great potentials in land-use classification and vegetation monitoring.