No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Super-resolution enhancement of Sentinel-2 image for retrieving LAI and chlorophyll content of summer corn
Abstract
Sentinel-2 satellite is a new generation of multi-spectral remote sensing technique with high spatial, temporal and spectral resolution. Especially, Sentinel-2 incorporates three red-edge bands with central wavelength at 705, 740 and 783 nm, which are very sensitive to vegetation changing, heath and variations. Unfortunately, their spatial resolution is only 20 m, which is lower comparably. This spatial resolution brings difficulties for mining the potential of Sentinel-2 image in vegetation monitoring. Therefore, we focus on enhancing the spatial resolution of Sentinel-2 red edge band images to 10m using the SupReME algorithm. Furthermore, the summer corn canopy leaf area index (LAI), leaves chlorophyll content (LCC) and canopy chlorophyll content (CCC) were retrieved by the linear and physical models for the corn growth monitoring purpose. The results showed that the spatial resolution of Sentinel-2 images had been enhanced to 10m from original 20m, and the estimation accuracy (EA) was over 97% for pixels planted by summer corn. Moreover, the accuracy of summer corn canopy LAI, LCC and CCC was improved respectively using enhanced Sentinel-2 images by SupReME method. During these three parameters retrieval, the red-edge bands or SWIR bands were introduced into optimal cost function and vegetation index which the accuracy of these models was high. The SupReME algorithm provides a valuable way for Sentinel-2 images enhancement, which is of great potential to mining Sentinel-2 images and multitude its application.
... In this study, the S2-MSI imagery was reconstructed at a spatial resolution of 10 m with Su-pReME to reduce the spatial ratio between UAV and satellite imagery. Since SupReME has been verified to well maintain the spectral and spatial consistency of S2-MSI imagery [31,42], we did not display the verification again. Then a relative radiometric normalization method was applied to mitigate the influence of sensor design. ...
... In this study, the S2-MSI imagery was reconstructed at a spatial resolution of 10 m with SupReME to reduce the spatial ratio between UAV and satellite imagery. Since SupReME has been verified to well maintain the spectral and spatial consistency of S2-MSI imagery [31,42], we did not display the verification again. Then a relative radiometric normalization method was applied to mitigate the influence of sensor design. ...
Accurate and continuous monitoring of crop growth is vital for the development of precision agriculture. Unmanned aerial vehicle (UAV) and satellite platforms have considerable complementarity in high spatial resolution (centimeter-scale) and fixed revisit cycle. It is meaningful to optimize the cross-platform synergy for agricultural applications. Considering the characteristics of UAV and satellite platforms, a spatio-temporal fusion (STF) framework of UAV and satellite imagery is developed. It includes registration, radiometric normalization, preliminary fusion, and reflectance reconstruction. The proposed STF framework significantly improves the fusion accuracy with both better quantitative metrics and visualized results compared with four existing STF methods with different fusion strategies. Especially for the prediction of object boundary and spatial texture, the absolute values of Robert’s edge (EDGE) and local binary pattern (LBP) decreased by a maximum of more than 0.25 and 0.10, respectively, compared with the spatial and temporal adaptive reflectance fusion model (STARFM). Moreover, the STF framework enhances the temporal resolution to daily, although the satellite imagery is discontinuous. Further, its application potential for winter wheat growth monitoring is explored. The daily synthetic imagery with UAV spatial resolution describes the seasonal dynamics of winter wheat well. The synthetic Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index 2 (EVI2) are consistent with the observations. However, the error in NDVI and EVI2 at boundary changes is relatively large, which needs further exploration. This research provides an STF framework to generate very dense and high-spatial-resolution remote sensing data at a low cost. It not only contributes to precision agriculture applications, but also is valuable for land-surface dynamic monitoring.
... Yet, the analysis revealed several deficiencies of the NIR/RE-model, raising the topics of temporal and spatial resolution. Apparently, the frequent statement that the GAI-course over the season can be well mapped with Sentinel-2 data [3,11,12,21,29,37,[40][41][42][43][44][45] is based on large differences and continuously increasing GAI-values between different sampling dates. Yet, the Sentinel-2 based single-date GAI-estimations are at some dates systematically biased and the true GAI-variation is in most cases considerably underestimated (Figure 4). ...
... This raises the question why so far, as to say based on the literature research carried out [3,11,12,21,29,37,[40][41][42][43][44][45], the impact of date-specific effects when working with Sentinel-2 data is not further discussed. The common practice of distributing the sampling dates over different BBCH stages (e.g., [11,29]) and combining data of different locations, years and cultivars [40,41,[43][44][45] might have disguised this problem so far-thereby, even if a certain shift between different satellite data acquisitions is detected, it cannot be clearly distinguished from other reasons, such as a starting VI-saturation at high GAIs, different VI-GAI-correlations at differing BBCH stages, or altering soil background. ...
An approach of exploiting and assessing the potential of Sentinel-2 data in the context of precision agriculture by using data from an unmanned aerial vehicle (UAV) is presented based on a four-year dataset. An established model for the estimation of the green area index (GAI) of winter wheat from a UAV-based multispectral camera was used to calibrate the Sentinel-2 data. Large independent datasets were used for evaluation purposes. Furthermore, the potential of the satellite-based GAI-predictions for crop monitoring and yield prediction was tested. Therefore, the total absorbed photosynthetic radiation between spring and harvest was calculated with satellite and UAV data and correlated with the final grain yield. Yield maps at the same resolution were generated by combining yield data on a plot level with a UAV-based crop coverage map. The best tested model for satellite-based GAI-prediction was obtained by combining the near-, infrared- and Red Edge-waveband in a simple ratio (R2 = 0.82, mean absolute error = 0.52 m2/m2). Yet, the Sentinel-2 data seem to depict average GAI-developments through the seasons, rather than to map site-specific variations at single acquisition dates. The results show that the lower information content of the satellite-based crop monitoring might be mainly traced back to its coarser Red Edge-band. Additionally, date-specific effects within the Sentinel-2 data were detected. Due to cloud coverage, the temporal resolution was found to be unsatisfactory as well. These results emphasize the need for further research on the applicability of the Sentinel-2 data and a cautious use in the context of precision agriculture.
... The traditional method of acquiring information on the location and severity of crop lodging relies on the measurement in field campaign, which is time-consuming, laborious, and subjective and cannot be performed on a regional scale. Fortunately, remote sensing provides a timely and reliable method, including high-spatial-resolution and multi-spectral features, for acquiring crop lodging information across large areas [4][5][6]. The earliest study can be traced back to the identification of winter wheat lodging using microcomputer-assisted video images [7]. ...
Efficient, fast, and accurate crop lodging monitoring is urgent for farmers, agronomists, insurance loss adjusters, and policymakers. This study aims to explore the potential of Chinese GF-1 PMS high-spatial-resolution images for corn lodging monitoring and to find a robust and efficient way to identify corn lodging accurately and efficiently. Three groups of image features and five machine-learning approaches are used for classifying non-lodged, moderately lodged, and severely lodged areas. Our results reveal that (1) the combination of spectral bands, optimized vegetation indexes, and texture features classify corn lodging with an overall accuracy of 93.81% and a Kappa coefficient of 0.91. (2) The random forest is an efficient, robust, and easy classifier to identify corn lodging with the F1-score of 0.95, 0.92, and 0.95 for non-lodged, moderately lodged, and severely lodged areas, respectively. (3) The GF-1 PMS image has great potential for identifying corn lodging on a regional scale.
... We selected the image of the growing seasons (June to September), this could be the reason for the high water content within the plant. The various studies also found that SWIR bands and indices are important for predicting forest structure, chlorophyll, and canopy cover (Bera et al., 2021;Derwin et al., 2020;Sothe, 2017;Wang et al., 2018;Zhang et al., 2019). Among the atmospheric indices, AFRI1.6 and AFRI2.1 were important for both Sentinel-2 and Landsat-8 based full models. ...
Quantifying canopy cover using Random Forest (RF) model’s appropriate tuning parameters value and sensor based predictor variables is always challenging, especially in fragmented dry deciduous forests. Therefore, this study was designed to compare the performances of Sentinel-2 and Landsat-8 based models using the RF model for predicting canopy cover, and assess variables relative importance and correlation. Sentinel-2 and Landsat-8 based bands and spectral indices were used as predictor variables. We compared different mtry, ntree and bag fraction values of RF model. R-square (R2) and root mean square error (RMSE) were used for comparing the model performance. The result showed that lowest RMSE value associated with default value (predictors/3) or more than default value of mtry, with bag fraction 0.3-0.7 for Sentinel-2 and 0.3-0.4 for Landsat-8 based full model. Model accuracy has increased and stabilized with increase number of tree, and received the lowest RMSE to ntree more than 1000. Except SWIR indices based model of Landsat-8, all other Landsat-8 based model’s accuracy was lesser compared to Sentinel-2 based models. Model accuracy of Sentinel-2 based full model (except red edge) was marginally better (R2 = 0.899, RMSE = 6.883 %) than Landsat-8 based full model (R2 = 0.886, RMSE = 7.089%). But with incorporation of red edge based indices, full model RMSE had decreased further 6.883% to 6.747%, and R2 had increased 0.899 to 0.918. Sentinel-2 based full model tendency to spread variable importance among more variables, but Landsat-8 based full model slightly tendency to concentrate variable importance with fewer variables. However, SWIR bands were the most important predictor variables and highly correlated with canopy cover. These findings can solve the choice of the parameters value of the RF model, and the use of the Sentinel-2 based model will be superior than Landsat-8 based model.
... The utility of S2-10m and S2-20m for various parameters is essential for the rapid assessment of the crop biophysical and biochemical parameters, without delays caused by additional pre-processing steps such as downsampling the S2-10m or upsampling the S2-20m spectral bands, and applying super-resolving techniques (Zhang et al. 2019), before retrieval; thus, the results from this study have operational significance. In our study, upsampling to 10 m caused 7 min and 17.787 s delay for a single Sentinel-2 tile consisting of width and height of 10,980 pixels on an IntelV R Core TM i7-8700 CPU and 64 GB RAM. ...
Sentinel-2 spectral configurations, S2-10m and S2-20m, were evaluated for retrieving essential crop biophysical and biochemical parameters and their effect on the performance of three machine learning regression algorithms (MLRAs) in two African semi-arid sites. The results were benchmarked against all spectral bands (S2-All). The results show that the S2-20m was more robust in retrieving Leaf Area Index (LAI) (RMSEcv: 0.58 m² m⁻², 0.47 m² m⁻²), while the S2-10m provided optimal retrievals Leaf Chlorophyll a + b (LCab) (RMSEcv: 6.89 µg cm⁻², 7.02 µg cm⁻²) for the two sites, respectively. In contrast, S2-20m performed better in retrieving Canopy Chlorophyll Content (CCC) in Bothaville to an RMSEcv of 35.65 µg cm⁻², while S2-10m yielded relatively lower uncertainties (RMSEcv of 26.84 µg cm⁻²) in Harrismith. Moreover, various MLRAs were sensitive to the various spectral configurations, and performance varied by site. GPR and XGBoost were more robust, and thus have the most potential for crop biophysical and biochemical parameter retrieval in both sites. Based on the benchmark results, the two configurations can be used independently. The results obtained here are relevant for the rapid development of essential crop biophysical and biochemical parameters for precision agriculture using Sentinel-2’s 10 m or 20 m bands, without the need for resampling.
... where SPAD is the value read from SPAD-502Plus and means the chlorophyll molar mass set at 907 g/mol [47]. ...
The current study aimed to determine the spatial transferability of eXtreme Gradient Boosting (XGBoost) models for estimating biophysical and biochemical variables (BVs), using Sentinel-2 data. The specific objectives were to: (1) assess the effect of different proportions of training samples (i.e., 25%, 50%, and 75%) available at the Target site () on the spatial transferability of the XGBoost models and (2) evaluate the effect of the Source site () (i.e., trained) model accuracy on the Target site (i.e., unseen) retrieval uncertainty. The results showed that the Bothaville () → Harrismith () Leaf Area Index (LAI) models required only fewer proportions, i.e., 25% or 50%, of the training samples to make optimal retrievals in the (i.e., RMSE: 0.61 m2 m–2; R2: 59%), while Harrismith () →Bothaville () LAI models required up to 75% of training samples in the to obtain optimal LAI retrievals (i.e., RMSE = 0.63 m2 m–2; R2 = 67%). In contrast, the chlorophyll content models for Bothaville () → Harrismith () required significant proportions of samples (i.e., 75%) from the to make optimal retrievals of Leaf Chlorophyll Content (LCab) (i.e., RMSE: 7.09 µg cm−2; R2: 58%) and Canopy Chlorophyll Content (CCC) (i.e., RMSE: 36.3 µg cm−2; R2: 61%), while Harrismith () →Bothaville () models required only 25% of the samples to achieve RMSEs of 8.16 µg cm−2 (R2: 83%) and 40.25 µg cm−2 (R2: 77%), for LCab and CCC, respectively. The results also showed that the source site model accuracy led to better transferability for LAI retrievals. In contrast, the accuracy of LCab and CCC source site models did not necessarily improve their transferability. Overall, the results elucidate the potential of transferable Machine Learning Regression Algorithms and are significant for the rapid retrieval of important crop BVs in data-scarce areas, thus facilitating spatially-explicit information for site-specific farm management.
... , and are similar or exceed LAI estimates from Sentinel-2 [e.g.61,62] and Landsat [e.g.[63][64][65]. ...
Satellite remote sensing has great potential to deliver on the promise of a data-driven agricultural revolution, with emerging space-based platforms providing spatiotemporal insights into precision-level attributes such as crop water use, vegetation health and condition and crop response to management practices. Using a harmonized collection of high-resolution Planet CubeSat, Sentinel-2, Landsat-8 and additional coarser resolution imagery from MODIS and VIIRS, we exploit a multi-satellite data fusion and machine learning approach to deliver a radiometrically calibrated and gap-filled time-series of daily leaf area index (LAI) at an unprecedented spatial resolution of 3 m. The insights available from such high-resolution CubeSat-based LAI data are demonstrated through tracking the growth cycle of a maize crop and identifying observable within-field spatial and temporal variations across key phenological stages. Daily LAI retrievals peaked at the tasseling stage, demonstrating their value for fertilizer and irrigation scheduling. An evaluation of satellite-based retrievals against field-measured LAI data collected from both rain-fed and irrigated fields shows high correlation and captures the spatiotemporal development of intra- and inter-field variations. Novel agricultural insights related to individual vegetative and reproductive growth stages were obtained, showcasing the capacity for new high-resolution CubeSat platforms to deliver actionable intelligence for precision agricultural and related applications.
... The spectral characteristics of different pest levels are analyzed for the Sentinel-2 image and the UAV images. For unifying the analyzing unit of spectral difference, the obtained Sentinel-2 image on August 18, 2019 in the study area is reconstructed with super resolution using SupReME, 39 and each band is unified to 10 m spatial resolution. After that, the spectral characteristics of Sentinel-2 image before and after super resolution reconstruction are analyzed, mainly in the building and corn planted area, which is as shown in Fig. 7. ...
Background: The timely, rapid, and accurate near real-time observations are urgent to monitor the damage of corn armyworm, because the rapid expansion of armyworm would lead to severe yield losses. Therefore, the potential of machine learning algorithms for identifying the armyworm infected areas automatically and accurately by multispectral Unmanned Aerial Vehicle (UAV) dataset is explored in this study. And the study area is in Beicuizhuang Village, Langfang City, Hebei Province, which is the main corn-producing area in the North China Plain.
Results: Firstly, we identified the optimal combination of image features by Gini-importance and the comparation of four kinds of machine learning methods including Random Forest (RF), Multilayer Perceptron (MLP), Naive Bayes Classifier (NB) and Support Vector Machine (SVM) was done. And RF was proved to be the most potential with the highest Kappa and OA of 0.9709 and 0.9850, respectively. Secondly, the armyworm infected areas and healthy corn areas were predicted by an optimized RF model in the UAV dataset, and the armyworm incidence levels were classified subsequently. Thirdly, the relationship between the spectral characteristics of different bands and pest incidence levels within the Sentinel-2 and UAV images were analyzed, and the B3 in UAV images and the B6 in Sentinel-2 image were less sensitive for armyworm incidence levels. So the Sentinel-2 image was used to monitor armyworm in two towns.
Conclusions: The optimized dataset and RF model are effective and reliable, which can be used for identifying the corn damage by armyworm using UAV images accurately and automatically in field-scale. This article is protected by copyright. All rights reserved.
... Six published optical VIs were calculated, according to the formulas listed in Table 3, which have been widely used in the estimation of key parameters of crops [36][37][38]. ...
The leaf area index (LAI) is of great significance for crop growth monitoring. Recently, unmanned aerial systems (UASs) have experienced rapid development and can provide critical data support for crop LAI monitoring. This study investigates the effects of combining spectral and texture features extracted from UAS multispectral imagery on maize LAI estimation. Multispectral images and in situ maize LAI were collected from test sites in Tongshan, Xuzhou, Jiangsu Province, China. The spectral and texture features of UAS multispectral remote sensing images are extracted using the vegetation indices (VIs) and the gray-level co-occurrence matrix (GLCM), respectively. Normalized texture indices (NDTIs), ratio texture indices (RTIs), and difference texture indices (DTIs) are calculated using two GLCM-based textures to express the influence of two different texture features on LAI monitoring at the same time. The remote sensing features are prescreened through correlation analysis. Different data dimensionality reduction or feature selection methods, including stepwise selection (ST), principal component analysis (PCA), and ST combined with PCA (ST_PCA), are coupled with support vector regression (SVR), random forest (RF), and multiple linear regression (MLR) to build the maize LAI estimation models. The results reveal that ST_PCA coupled with SVR has better performance, in terms of the VIs + DTIs (R2 = 0.876, RMSE = 0.239) and VIs + NDTIs (R2 = 0.877, RMSE = 0.236). This study introduces the potential of different texture indices for maize LAI monitoring and demonstrates the promising solution of using ST_PCA to realize the combining of spectral and texture features for improving the estimation accuracy of maize LAI.
... As the most efficient method of obtaining the land surface information in regional areas [1,2], space-based optical remote sensing devices have been significantly improved in the last few decades. The remote sensing image analysis methods based on pixels have promoted the continuous progress of image analysis techniques and theory. ...
High-spatial-resolution images play an important role in land cover classification, and object-based image analysis (OBIA) presents a good method of processing high-spatial-resolution images. Segmentation, as the most important premise of OBIA, significantly affects the image classification and target recognition results. However, scale selection for image segmentation is difficult and complicated for OBIA. The main challenge in image segmentation is the selection of the optimal segmentation parameters and an algorithm that can effectively extract the image information. This paper presents an approach that can effectively select an optimal segmentation scale based on land object average areas. First, 20 different segmentation scales were used for image segmentation. Next, the classification and regression tree model (CART) was used for image classification based on 20 different segmentation results, where four types of features were calculated and used, including image spectral bands value, texture value, vegetation indices, and spatial feature indices, respectively. WorldView-3 images were used as the experimental data to verify the validity of the proposed method for the selection of the optimal segmentation scale parameter. In order to decide the effect of the segmentation scale on the object area level, the average areas of different land objects were estimated based on the classification results. Experiments based on the multi-scale segmentation scale testify to the validity of the land object’s average area-based method for the selection of optimal segmentation scale parameters. The study results indicated that segmentation scales are strongly correlated with an object’s average area, and thus, the optimal segmentation scale of every land object can be obtained. In this regard, we conclude that the area-based segmentation scale selection method is suitable to determine optimal segmentation parameters for different land objects. We hope the segmentation scale selection method used in this study can be further extended and used for different image segmentation algorithms.
... Remote sensing is an efficient way to capture the land surface information quickly in regional areas [8,9]. The crop residue cover estimation and the conservative tillage monitoring based on remote sensing data have become a topic of significant interest to researchers [10,11]. ...
Black soil is one of the most productive soils with high organic matter content. Crop residue covering is important for protecting black soil from alleviating soil erosion and increasing soil organic carbon. Mapping crop residue covered areas accurately using remote sensing images can monitor the protection of black soil in regional areas. Considering the inhomogeneity and randomness, resulting from human management difference, the high spatial resolution Chinese GF-1 B/D image and developed MSCU-net+C deep learning method are used to mapping corn residue covered area (CRCA) in this study. The developed MSCU-net+C is joined by a multiscale convolution group (MSCG), the global loss function, and Convolutional Block Attention Module (CBAM) based on U-net and the full connected conditional random field (FCCRF). The effectiveness of the proposed MSCU-net+C is validated by the ablation experiment and comparison experiment for mapping CRCA in Lishu County, Jilin Province, China. The accuracy assessment results show that the developed MSCU-net+C improve the CRCA classification accuracy from IOUAVG = 0.8604 and KappaAVG = 0.8864 to IOUAVG = 0.9081 and KappaAVG = 0.9258 compared with U-net. Our developed and other deep semantic segmentation networks (MU-net, GU-net, MSCU-net, SegNet, and Dlv3+) improve the classification accuracy of IOUAVG/KappaAVG with 0.0091/0.0058, 0.0133/0.0091, 0.044/0.0345, 0.0104/0.0069, and 0.0107/0.0072 compared with U-net, respectively. The classification accuracies of IOUAVG/KappaAVG of traditional machine learning methods, including support vector machine (SVM) and neural network (NN), are 0.576/0.5526 and 0.6417/0.6482, respectively. These results reveal that the developed MSCU-net+C can be used to map CRCA for monitoring black soil protection.
... Sentinel-2 has four red-edge bands (B5, B6, B7, and B8A), which provide important data for complex crop classification. B5 and B6 are particularly useful for red-edge position, B7 for calculating the inversion of the leaf area index (LAI) [58], and B8A is sensitive to LAI, chlorophyll, and biomass [59]. The data have been widely used in agricultural monitoring [60], ecological assessments [61], and land cover change analyses [62]. ...
Accurate temporal land use mapping provides important and timely information for decision making for large-scale management of land and crop production. At present, temporal land cover and crop classifications within a study area have neglected the differences between subregions. In this paper, we propose a classification rule by integrating the terrain, time series characteristics, priority, and seasonality (TTPSR) with Sentinel-2 satellite imagery. Based on the time series of Normalized Difference Water Index (NDWI) and Vegetation Index (NDVI), a dynamic decision tree for forests, cultivation, urban, and water was created in Google Earth Engine (GEE) for each subregion to extract cultivated land. Then, with or without this cultivated land mask data, the original classification results for each subregion were completed based on composite image acquisition with five vegetation indices using Random Forest. During the post-reclassification process, a 4-bit coding rule based on terrain, type, seasonal rhythm, and priority was generated by analyzing the characteristics of the original results. Finally, statistical results and temporal mapping were processed. The results showed that feature importance was dominated by B2, NDWI, RENDVI, B11, and B12 over winter, and B11, B12, NDBI, B2, and B8A over summer. Meanwhile, the cultivated land mask improved the overall accuracy for multicategories (seven to eight and nine to 13 during winter and summer, respectively) in each subregion, with average ranges in the overall accuracy for winter and summer of 0.857–0.935 and 0.873–0.963, respectively, and kappa coefficients of 0.803–0.902 and 0.835–0.950, respectively. The analysis of the above results and the comparison with resampling plots identified various sources of error for classification accuracy, including spectral differences, degree of field fragmentation, and planting complexity. The results demonstrated the capability of the TTPSR rule in temporal land use mapping, especially with regard to complex crops classification and automated post-processing, thereby providing a viable option for large-scale land use mapping.
... To quantify the performance of the super-resolution algorithm, Pearson's correlation coefficient (R) [44], the root mean squared error (RMSE) [45] and estimation accuracy (EA) [46] were calculated for each band as follows (Equations (7) and (8)): ...
In the application of quantitative remote sensing in water quality monitoring, the existence of mixed pixels greatly affects the accuracy of water quality parameter inversion, especially for narrow inland rivers. Improving the image spatial resolution and weakening the interference of mixed pixels in the image are some of the urgent problems to be solved in the study of water quality monitoring of medium- and small-sized inland rivers. We processed Sentinel-2 multispectral images using the super-resolution algorithm and generated a set of 10 m spatial resolution images with basically unchanged reflection characteristics. Both qualitative and quantitative evaluation results show that the super-resolution algorithm can weaken the influence of mixed pixels while maintaining spectral invariance. Before the application of the super-resolution algorithm, the inversion accuracy of water quality parameters in this study were as follows: for NH3-N, the R2 was 0.61, the root mean squared error (RMSE) was 0.177 and the mean absolute percentage error (MAPE) was 29.33%; for Chemical Oxygen Demand (COD), the R2 was 0.26, the RMSE was 0.756 and the MAPE was 4.62%; for Total Phosphorus (TP), the R2 was 0.69, the RMSE was 0.032 and the MAPE was 30.58%. After the application of the super-resolution algorithm, the inversion accuracy of water quality parameters in this study were as follows: for NH3-N, the R2 was 0.67, the RMSE was 0.161 and the MAPE was 25.88%; for COD, the R2 was 0.53, the RMSE was 0.546 and the MAPE was 3.36%; for TP, the R2 was 0.60, the RMSE was 0.034 and the MAPE was 24.28%. Finally, the spatial distribution of NH3-N, COD and TP was obtained by using a machine learning model. The results showed that the application of the super-resolution algorithm can effectively improve the retrieval accuracy of NH3-N, COD and TP, which illustrates the application potential of the super-resolution algorithm in water quality remote sensing quantitative monitoring.
... LAI is widely used in crop growth monitoring, land-surface process simulation, and global change studies (Duan et al., 2019;Xie et al., 2019;Zhang et al., 2019) Presently, most studies utilize NDVI or FVC as tools to analyze the growth of vegetation, while LAI is seldom considered in such analyses (He et al., 2017;Yang et al., 2018;Zhang et al., 2016). Considering that LAI has an obvious seasonal variation, the annual maximum LAI value was used to represent vegetation growth conditions. ...
The Three-North Shelter Forest Program (TNSFP) region of China is an important ecological region covering more than 42.4% of China’s land area. Several ecological restoration projects have been implemented in this region, but evaluation work is relatively limited partly due to the selection of suitable remote sensing products. In this study, leaf area index is proposed as a suitable indicator to evaluate the ecological restoration situation as it is an indicator reflects the growth of green vegetations. The Global Land Surface Satellite (GLASS) LAI dataset and land cover dataset from 2000 to 2015 were adopted and analyzed in this study. The forest cover area slightly increased, while the grassland area decreased. Although large-scale forestation movements were carried out across the entire TNSFP region, the forest growth conditions mainly improved in regions with an annual precipitation greater than 400 mm. Furthermore, LAI-precipitation correlations were very high (r greater than 0.6) in the regions across 400 mm isohyets, but lower (r < 0.4) far away from 400 mm isohyets. In relatively arid regions, forest growth did not show an obvious increase trend and sometimes a decrease trend. The increase in the LAI of these regions was mainly due to the restoration of grassland. Therefore, although LAI in most of the TNSFP region showed increasing trends, it is still not sufficient to state that these trends were caused by forestation projects in China.
... The land surface water index (LSWI) more accurately classify different crops, especially paddy rice [41]. Other indicators used in our research included the red-edge parameter (REP) and the blue-green (B-G) ratio [42,43]. We also designed a new indicator called the normalized difference rice index (NDRI), which can better identify the paddy rice on the Northeast Plain. ...
Large-scale, high-resolution mapping of crop patterns is useful for the assessment of food security and agricultural sustainability but is still limited. This study attempted to establish remote sensing-based crop classification models for specific cropping systems using the decision trees method and monitored the distribution of the major crop species using Sentinel-2 satellites (10 m) in 2017. The results showed that the cropping areas of maize, rice, and soybean on the Northeast China Plain were approximately 12.1, 6.2, and 7.4 million ha, respectively. The cropping areas of winter wheat and summer maize on the North China Plain were 13.4 and 16.9 million ha, respectively. The cropping areas of wheat, rice, and rape on the middle-lower Yangtze River plain were 2.2, 6.4 and 1.3 million ha, respectively. Estimated images agreed well with field survey data (average overall accuracy = 94%) and the national agricultural census data (R 2 = 0.78). This indicated the applicability of the Sentinel-2 satellite data for large-scale, high-resolution crop mapping in China. We intend to update the crop mapping datasets annually and hope to guide the adjustment and optimization of the national agricultural structure.
... The spectral indices captured by hyperspectral sensors are considered to be quantitative indicators of vegetation vigor (Hunt et al., 2011;Jay et al., 2017;Peng et al., 2017;Upreti et al., 2019), and the continuous narrow bands have the potential to identify bands sensitive to specific crop parameters (Gitelson et al., 2003). Owing to these attributes, hyperspectral remote sensing (RS) has been broadly applied in the estimation or retrieval of crop chlorophyll content (Sonobe et al., 2018;Zhang et al., 2019;Ali and Imran, 2020). However, precision agriculture requires the use of large-scale crop data at high temporal and spatial resolutions to enable timely, accurate, and cost-effective within-field monitoring (Kross et al., 2015). ...
Leaf chlorophyll content (LCC) is a crucial indicator of nutrition in crop plants and can be applied to assess the adequacy of nitrogen (N) fertilizer for crops while reducing N losses to farmland. This study estimated the LCC of maize and wheat, and comprehensively examined the effects of the spectral information and spatial scale of unmanned aerial vehicle (UAV) imagery, and the effects of phenotype and phenology on LCC estimation. A Cubert S185 hyperspectral camera onboard a DJI M600 Pro was used to conduct six flight missions over a long-term experimental field with five N applications (0, 70, 140, 210, and 280 kg N ha−1) and two irrigation levels (60% and 80% field water capacity) during the growing seasons of wheat and maize in 2019. Four regression algorithms, that is, multi-variable linear regression, random forest, backpropagation neural network, and support vector machine, were used for modeling. Leaf, canopy, and hybrid scale hyperspectral variables (H-variables) were used as inputs for the statistical LCC models. Optimal H-variables for modeling were determined by Pearson correlation filtering followed by a recursive feature elimination procedure. The results showed that (1) H-variables at the canopy- and leaf-scales were appropriate for wheat and maize LCC estimation, respectively; (2) the robustness of LCC estimation was in the order of the flowering stage > heading stage > grain filling stage for wheat and early grain filling stage > flowering stage > jointing stage for maize; (3) the reflectance of the red edge, green, and blue bands were the most important inputs for LCC modeling, and the optimal vegetation indices differed for the various growth stages and crops; and (4) all four algorithms maintained an acceptable accuracy with respect to LCC estimation, although random forest and support vector machine were slightly better. This study is valuable for the design of appropriate schemes for the spectral and scale issues of UAV sensors for LCC estimation regarding specific crop phenotype and phenology periods, and further boosts the applications of UAVs in precision agriculture.
... In light of its free availability, world-wide coverage, revisit frequency and, not least, its above remarked wide applicability, several research teams have proposed solutions to super-resolve Sentinel-2 images, rising 20 m and/or 60 m bands up to 10 m resolution. Besides, several works testify the advantage of using super-resolved S2 images in several applications such as water mapping [11], fire detection [12], urban mapping [13], and vegetation monitoring [14]. ...
Images provided by the ESA Sentinel-2 mission are rapidly becoming the main source of information for the entire remote sensing community, thanks to their unprecedented combination of spatial, spectral and temporal resolution, as well as their associated open access policy. Due to a sensor design trade-off, images are acquired (and delivered) at different spatial resolutions (10, 20 and 60 m) according to specific sets of wavelengths, with only the four visible and near infrared bands provided at the highest resolution (10 m). Although this is not a limiting factor in general, many applications seem to emerge in which the resolution enhancement of 20 m bands may be beneficial, motivating the development of specific super-resolution methods. In this work, we propose to leverage Convolutional Neural Networks (CNNs) to provide a fast, upscalable method for the single-sensor fusion of Sentinel-2 (S2) data, whose aim is to provide a 10 m super-resolution of the original 20 m bands. Experimental results demonstrate that the proposed solution can achieve better performance with respect to most of the state-of-the-art methods, including other deep learning based ones with a considerable saving of computational burden.
During the growth season, jujube trees are susceptible to infestation by the leaf mite, which reduces the fruit quality and productivity. Traditional monitoring techniques for mites are time-consuming, difficult, subjective, and result in a time lag. In this study, the method based on a particle swarm optimization (PSO) algorithm extreme learning machine for estimation of leaf chlorophyll content (SPAD) under leaf mite infestation in jujube was proposed. Initially, image data and SPAD values for jujube orchards under four severities of leaf mite infestation were collected for analysis. Six vegetation indices and SPAD value were chosen for correlation analysis to establish the estimation model for SPAD and the vegetation indices. To address the influence of colinearity between spectral bands, the feature band with the highest correlation coefficient was retrieved first using the successive projection algorithm. In the modeling process, the PSO correlation coefficient was initialized with the convergent optimal approximation of the fitness function value; the root mean square error (RMSE) of the predicted and measured values was derived as an indicator of PSO goodness-of-fit to solve the problems of ELM model weights, threshold randomness, and uncertainty of network parameters; and finally, an iterative update method was used to determine the particle fitness value to optimize the minimum error or iteration number. The results reflected that significant differences were observed in the spectral reflectance of the jujube canopy corresponding with the severity of leaf mite infestation, and the infestation severity was negatively correlated with the SPAD value of jujube leaves. The selected vegetation indices NDVI, RVI, PhRI, and MCARI were positively correlated with SPAD, whereas TCARI and GI were negatively correlated with SPAD. The accuracy of the optimized PSO-ELM model ( R ² = 0.856, RMSE = 0.796) was superior to that of the ELM model alone ( R ² = 0.748, RMSE = 1.689). The PSO-ELM model for remote sensing estimation of relative leaf chlorophyll content of jujube shows high fault tolerance and improved data-processing efficiency. The results provide a reference for the utility of UAV remote sensing for monitoring leaf mite infestation of jujube.
Multi-sensor fusion provides an effective way for applications requiring remote sensing data with high spatiotemporal resolution. Especially for agricultural areas with complex planting structures and rapid changes in crop phenology, more detailed and dense time-series remote sensing data are necessary. The Sentinel-2 Multispectral Imager (S2-MSI) sensor with high spatial resolution (10–60 m) and temporal resolution (5–10 days) plays a key role in spatiotemporal fusion. But the inconsistent spatial resolution of the various bands hinders its potential application at 10 m resolution, and the multiple available fine images it provides are not fully utilized for spatiotemporal fusion. It is worth exploring how to maximize the spatial and temporal resolution of S2-MSI images to help improve the effect of spatiotemporal fusion and the dynamic monitoring of rapid crop growth. In this research, a new spatiotemporal fusion (STF) framework is developed to fuse the S2-MSI image (10 m) enhanced by Super-Resolution for multispectral Multiresolution Estimation (SupReME) algorithm and MODIS image (460 m) with a large spatial ratio (46). The proposed fusion method in the new STF framework combines the existing STF methods with Consistent Adjustment of the Climatology to Actual Observations (CACAO) algorithm, abbreviated as CA-STF. The accuracy of the fused reflectance and its capability for dynamic LAI monitoring were tested in Daman Superstation of Heihe watershed. The results indicate that: (1) the new STF framework is competent to fuse multi-source images with a ratio of 46 and outperforms the existing STF methods for both near-real-time and post-growth applications; (2) the proposed CA-STF method improves the fusion accuracy and spatial details even if only two S2-MSI images are available, especially for post-growth applications; (3) the vegetation indices (VIs) calculated from the fused images by the new STF framework provide a better correlation with LAINet measurements and improve dynamic LAI monitoring in accuracy and spatial details. This study proposes a framework to maximize the spatial and temporal resolution of S2-MSI images for spatiotemporal fusion. The synthetic daily time-series images with a high resolution of 10 m will have great potential for monitoring the dynamic changes of the land surface.
ESTIMATIVA DO ÍNDICE DE ÁREA FOLIAR E DA FRAÇÃO DE COBERTURA DO SOLO NAS CULTURAS DE MILHO E SOJA USANDO NDVI1 LEORNARDO CHECHI2; MIRTA TERESINHA PETRY3; ZANANDRA BOFF DE OLIVEIRA4; MAX KLEBER LAURENTINO DANTAS5; CLARISSA MORAES DA SILVA6 E ANDRESSA FUZER GONÇALVES7 1 Parte do trabalho de dissertação de Mestrado do primeiro autor - Programa de Pós-Graduação em Engenharia Agrícola – UFSM 2 Eng. Agrônomo, Mestre em Engenharia Agrícola, Programa de Pós-Graduação em Engenharia Agrícola, UFSM, Avenida Roraima, nº1000, Bairro Camobi, Santa Maria, RS, Brasil, CEP: 97195-000, leonardo.chechi@gmail.com 3 Eng. Agrônoma, Doutora, Professora do Departamento de Engenharia Rural, UFSM, Avenida Roraima, nº1000, Bairro Camobi, Santa Maria, RS, Brasil, CEP: 97195-000, mirta.petry@gmail.com 4 Eng. Agrícola, Doutora, Coordenadoria acadêmica Curso de Engenharia Agrícola, Campus da UFSM de Cachoeira do Sul. Rod. Taufik Germano, n.3013, Bairro Passo D’Areia, Cachoeira do Sul, RS, Brasil, CEP: 96503-205, zanandraboff@gmail.com. 5 Eng. Agrônomo, Doutor em Ciência do Solo, UFSM, Avenida Roraima, nº1000, Bairro Camobi, Santa Maria, RS, Brasil, CEP: 97195-000, maxdantas22@gmail.com. 6 Eng. Agrícola, Aluna de Mestrado do Programa de Pós-Graduação em Engenharia Agrícola, UFSM, Avenida Roraima, nº1000, Bairro Camobi, Santa Maria, RS, Brasil, CEP: 97195-000, clarissamoraes37@outlook.com. 7 Acadêmica Curso de Agronomia, UFSM, Avenida Roraima, nº1000, Bairro Camobi, Santa Maria, RS, Brasil, CEP: 97195-000, dessa.fuzer@gmail.com. 1 RESUMO Técnicas de sensoriamento remoto são utilizadas para avaliar mudanças na paisagem, como a fenologia, índice de área foliar (IAF), altura de plantas e fração de cobertura do solo (fc). Os índices de vegetação (VI) têm sido relacionados às características biofísicas das culturas, como o IAF e a fc. Assim, o NDVI (Normalized Difference Vegetation Index) tem sido utilizado para verificar a dinâmica da vegetação durante o ciclo de cultivo. Nesse estudo, o NDVI foi usado para estimar a fc e o IAF das culturas soja e milho, visando monitorar o estado atual da vegetação para avaliações das necessidades hídricas das culturas. Observações à campo do IAF e fc foram realizadas em intervalos de aproximadamente 8 dias. O NDVI foi derivado de imagens do satélite Sentinel (sensores 2A e 2B) e processado na biblioteca open-source do Google Earth Engine. Observou-se elevado ajuste entre os valores observados e simulados, com valores de b0 próximos a 1,00, R² > 0,99 e RMSE variando de 0,02 a 0,05 para a fc e de 0,29 a 0,61 para o IAF, indicando que os modelos propostos foram eficientes no monitoramento dessas variáveis biofísicas. Palavras-chave: Sentinel, Google Engine, índices de vegetação, IAF. CHECHI, L., PETRY, M.T., OLIVEIRA, Z. B., DANTAS, M. K. L., SILVA, C. M.; GONÇALVES, A. F. ESTIMATION OF LEAF AREA INDEX AND FRACTON OF GROUND COVER OF CORN AND SOYBEAN CROPS USING NDVI 2 ABSTRACT Remote sensing techniques are used to assess changes in the landscape, such as phenology, leaf area index (LAI), plant height, and the fraction of ground cover (fc). Vegetation indices (VI) have been related to biophysical characteristics of crops, such as LAI and fc. Thus, the NDVI (Normalized Difference Vegetation Index) has been used to monitor vegetation dynamics throughout the crop development cycle. In this study, the NDVI was used to estimate the fc and LAI of soybeans and corn crops, aiming to monitor the actual condition of the vegetation for crop water requirement assessments. Field observations of the LAI and fc were carried out at intervals of approximately eight days. The NDVI used to estimate the fc and LAI were derived from images of the Sentinel satellite (sensors 2A and 2B) and processed in the open-source Google Earth Engine library. A excellent fit between the observed and simulated values was observed, with values of b0 close to 1.00, R² > 0.99, and RMSE ranging from 0.02 to 0.05 for fc and from 0.29 to 0.61 for LAI, indicating that the proposed models were efficient in the monitoring of these biophysical variables. Keywords: Sentinel, Google Engine, vegetation indices, LAI.
Appropriate modeling methods and feature selection algorithms must be selected to improve the accuracy of early and mid-term remote sensing detection of wheat stripe rust. In the current study, we explored the effectiveness of the random forest (RF) algorithm combined with the extreme gradient boosting (XGboost) method for early and mid-term wheat stripe rust detection based on the vegetation indices extracted from canopy level hyperspectral measurements. Initially, 21 vegetation indices that were related to the early and mid-term winter wheat stripe rust were calculated on the basis of canopy level hyperspectral reflectance. Subsequently, the optimal vegetation index combination for disease detection was determined using correlation analysis (CA) combined with RF algorithms. Then, the disease severity detection model of early and mid-term winter wheat stripe rust was constructed using XGBoost method based on the optimal vegetation index combination. For the evaluation and comparison of the initial results, three commonly used classification methods, namely, RF, backpropagation neural network (BPNN), and support vector machine (SVM), were utilized. The vegetation index combinations determined by the single CA algorithm were also used to construct detection models. Compared with the detection models based on the vegetation index combination obtained using the single CA algorithm, the overall accuracy of the four detection models based on the optimal vegetation index combination based on CA combined with RF algorithms increased by 16.1% (XGBoost), 9.7% (RF), 8.1% (SVM), and 8.1% (BPNN). Among the eight models, the XGBoost detection model based on the optimal vegetation index combination using CA combined with RF algorithms, CA-RF-XGBoost, achieved the highest overall accuracy of 87.1% and the highest kappa coefficient of 0.798. Our results indicate that the RF combined with XGBoost can improve the detection accuracy of early and mid-term winter wheat stripe rust effectively at canopy scale.
Satellite remote sensing has great potential to deliver on the promise of a data-driven agricultural revolution, with emerging space-based platforms providing spatiotemporal insights into precision-level attributes such as crop water use, vegetation health and condition and crop response to management practices. Using a harmonized collection of high-resolution Planet CubeSat, Sentinel-2, Landsat-8 and additional coarser resolution imagery from MODIS and VIIRS, we exploit a multi-satellite data fusion and machine learning approach to deliver a radiometrically calibrated and gap-filled time-series of daily leaf area index (LAI) at an unprecedented spatial resolution of 3 m. The insights available from such high-resolution CubeSat-based LAI data are demonstrated through tracking the growth cycle of a maize crop and identifying observable within-field spatial and temporal variations across key phenological stages. Daily LAI retrievals peaked at the tasseling stage, demonstrating their value for fertilizer and irrigation scheduling. An evaluation of satellite-based retrievals against field-measured LAI data collected from both rain-fed and irrigated fields shows high correlation and captures the spatiotemporal development of intra- and inter-field variations. Novel agricultural insights related to individual vegetative and reproductive growth stages were obtained, showcasing the capacity for new high-resolution CubeSat platforms to deliver actionable intelligence for precision agricultural and related applications.
The aim of this study was to assess the utility of Sentinel-2 images in the monitoring of the fractional vegetation cover (FVC) of rainfed alfalfa in semiarid areas such as that of Bardenas Reales in Spain. FVC was sampled in situ using 1 m2 surfaces at 172 points inside 18 alfalfa fields from late spring to early summer in 2017 and 2018. Different vegetation indices derived from a series of Sentinel-2 images were calculated and were then correlated with the FVC measurements at the pixel and parcel levels using different types of equations. The results indicate that the normalized difference vegetation index (NDVI) and FVC were highly correlated at the parcel level (R2 = 0.712), whereas the correlation at the pixel level remained moderate across each of the years studied. Based on the findings, another 29 alfalfa plots (28 rainfed; 1 irrigated) were remotely monitored operationally for 3 years (2017–2019), revealing that location and weather conditions were strong determinants of alfalfa growth in Bardenas Reales. The results of this study indicate that Sentinel-2 imagery is a suitable tool for monitoring rainfed alfalfa pastures in semiarid areas, thus increasing the potential success of pasture management.
All ground-based estimations of leaf area index (LAI) of Moso bamboo canopies are currently conducted based on indirect remote sensing methods. However, the relatively small values of LAI estimated by previous studies conflict with the expected values of such extremely dense canopies of Moso bamboo. This is the first attempt to accurately estimate the LAI of Moso bamboo canopies using an allometric model based on destructive measurements. The results indicate that (1) LAI of Moso bamboo canopies range was 6.7–30.6 m²·m⁻², which is clearly higher than the range 2.2–6.5 m²·m⁻² estimated by previous studies; (2) there is a strong linear relationship between LAI and crown density (R² = 0.947, RMSE = 1.343); (3) LAI is largely underestimated using the digital hemispherical photography (DHP) because of the overestimation of clumping index; and (4) there is a strong exponential relationship between LAI and effective leaf area (Le) estimated using DHP (R² = 0.734, RMSE = 3.011). Based on the results, three methods are recommended for LAI estimations of Moso bamboo canopies using the allometric relationship, the empirical relationship with crown density, and the empirical relationship with Le.
Sea surface temperature (SST) can be measured from space using infrared sensors on Earth-observing satellites. However, the tradeoff between spatial resolution and swath size (and hence revisit time) means that SST products derived from remote sensing measurements commonly only have a moderate resolution (>1 km). In this article, we adapt the design of a super-resolution neural network architecture [specifically very deep super-resolution (VDSR)] to enhance the resolution of both top-of-atmosphere thermal images of sea regions and bottom-of-atmosphere SST images by a factor of 5. When tested on an unseen dataset, the trained neural network yields thermal images that have an RMSE
$2-3\times $
smaller than interpolation, with a 6–9 dB improvement in PSNR. A major contribution of the proposed neural network architecture is that it fuses optical and thermal images to propagate the high-resolution information present in the optical image to the restored thermal image. To illustrate the potential benefits of using super-resolution (SR) in the context of oceanography, we present super-resolved SST images of a gyre and an ocean front, revealing details and features otherwise poorly resolved by moderate resolution satellite images.
Key message
High-throughput phenotyping (HTP) is in its infancy for deployment in large-scale breeding programmes. With the ability to measure correlated traits associated with physiological ideotypes, in-field phenotyping methods are available for screening of abiotic stress responses. As cropping environments become more hostile and unpredictable due to the effects of climate change, the need to characterise variability across spatial and temporal scales will become increasingly important. The sensor technologies that have enabled HTP from macroscopic through to satellite sensors may also be utilised here to complement spatial characterisation using envirotyping, which can improve estimations of genotypic performance across environments by better accounting for variation at the plot, trial and inter-trial levels.
AbstractClimate change is leading to increased variation at all physical and temporal scales in the cropping environment. Maintaining yield stability under circumstances with greater levels of abiotic stress while capitalising upon yield potential in good years, requires approaches to plant breeding that target the physiological limitations to crop performance in specific environments. This requires dynamic modelling of conditions within target populations of environments, GxExM predictions, clustering of environments so breeding trajectories can be defined, and the development of screens that enable selection for genetic gain to occur. High-throughput phenotyping (HTP), combined with related technologies used for envirotyping, can help to address these challenges. Non-destructive analysis of the morphological, biochemical and physiological qualities of plant canopies using HTP has great potential to complement whole-genome selection, which is becoming increasingly common in breeding programmes. A range of novel analytic techniques, such as machine learning and deep learning, combined with a widening range of sensors, allow rapid assessment of large breeding populations that are repeatable and objective. Secondary traits underlying radiation use efficiency and water use efficiency can be screened with HTP for selection at the early stages of a breeding programme. HTP and envirotyping technologies can also characterise spatial variability at trial and within-plot levels, which can be used to correct for spatial variations that confound measurements of genotypic values. This review explores HTP for abiotic stress selection through a physiological trait lens and additionally investigates the use of envirotyping and EC to characterise spatial variability at all physical scales in METs.
Canopy cover is an important biophysical indicator of forested habitats, but quantifying and monitoring with the accurate and cost-efficient method is always difficult in fragmented canopy cover areas. Therefore, this study was designed to explore the comparative performance of linear regression, polynomial regression and generalized additive model (GAM), and assess the important predictors for canopy cover estimation in the dry deciduous forest of West Bengal. We used the Sentinel-2 based individual bands and spectral indices as predictor variables from the different composition of imagery in the growing season (June - September) between 2018 and 2020. Canopy cover of 667 plots were measured using collect earth online (CEO) from high resolution imagery. Univariate and multiple models were fitted and assessed by 10 fold cross-validation. Root mean square error (RMSE) and the coefficient of determination (R2) was used for the evaluation of model performance. The result showed that univariate models using GAM were fitted better followed polynomial regression and linear regression model. SWIR bands and SWIR bands based indices were fitted well than other groups of variables. Individual bands based multiple prediction models performed better followed by NIR, SWIR, atmospheric and red edge indices based multiple prediction models for multiple linear regression, multiple polynomial regression and multiple GAM. Multiple GAM performed better (RMSE = 0.163, R2 = 0.777) than multiple polynomial regression (RMSE = 0.167, R2 = 0.772) and multiple linear regression model (RMSE = 0.182, R2 = 0.721). Although, the difference of performance between multiple polynomial regression model and multiple GAM was negligible. But in respect to prediction range, and overestimation and underestimation multiple GAM was superior to multiple polynomial regression model. The study demonstrates that the potential utility of GAM for canopy cover estimation and GAM can be superior to linear and polynomial regression model using suitable predictor variables.
Keywords
Sentinel-2; Google earth engine; Multiple prediction model; Generalized additive model; Canopy cover; Vegetation indices
Yellow rust is a worldwide disease that poses a serious threat to the safety of wheat production. Numerous studies on near-surface hyperspectral remote sensing at the leaf scale have achieved good results for disease monitoring. The next step is to monitor the disease at the field scale, which is of great significance for disease control. In our study, an unmanned aerial vehicle (UAV) equipped with a hyperspectral sensor was used to obtain hyperspectral images at the field scale. Vegetation indices (VIs) and texture features (TFs) extracted from the UAV-based hyperspectral images and their combination were used to establish partial least-squares regression (PLSR)-based disease monitoring models in different infection periods. In addition, we resampled the original images with 1.2 cm spatial resolution to images with different spatial resolutions (3 cm, 5 cm, 7 cm, 10 cm, 15 cm, and 20 cm) to evaluate the effect of spatial resolution on disease monitoring accuracy. The findings showed that the VI-based model had the highest monitoring accuracy (R2 = 0.75) in the mid-infection period. The TF-based model could be used to monitor yellow rust at the field scale and obtained the highest R2 in the mid- and late-infection periods (0.65 and 0.82, respectively). The VI-TF-based models had the highest accuracy in each infection period and outperformed the VI-based or TF-based models. The spatial resolution had a negligible influence on the VI-based monitoring accuracy, but significantly influenced the TF-based monitoring accuracy. Furthermore, the optimal spatial resolution for monitoring yellow rust using the VI-TF-based model in each infection period was 10 cm. The findings provide a reference for accurate disease monitoring using UAV hyperspectral images.
Timely monitoring of crop lands is important in order to make agricultural activities more sustainable, as well as ensuring food security. The use of Earth Observation (EO) data allows crop monitoring at a range of spatial scales, but can be hampered by limitations in the data. Crop growth modelling, on the other hand, can be used to simulate the physiological processes that result in crop development. Data assimilation (DA) provides a way of blending the monitoring properties of EO data with the predictive and explanatory abilities of crop growth models. In this paper, we first provide a critique of both the advantages and disadvantages of both EO data and crop growth models. We use this to introduce a solid and robust framework for DA, where different DA methods are shown to be derived from taking different assumptions in solving for the a posteriori probability density function (pdf) using Bayes' rule. This treatment allows us to provide some recommendation on the choice of DA method for particular applications. We comment on current computational challenges in scaling DA applications to large spatial scales. Future areas of research are sketched, with an emphasis on DA as an enabler for blending different observations, as well as facilitating different approaches to crop growth models. We have illustrated this review with a large number of examples from the literature.
Upcoming satellite hyperspectral sensors require powerful and robust methodologies for making optimum use of the rich spectral data. This paper reviews the widely applied coupled PROSPECT and SAIL radiative transfer models (PROSAIL), regarding their suitability for the retrieval of biophysical and biochemical variables in the context of agricultural crop monitoring. Evaluation was carried out using a systematic literature review of 281 scientific publications with regard to their (i) spectral exploitation, (ii) vegetation type analyzed, (iii) variables retrieved, and (iv) choice of retrieval methods. From the analysis, current trends were derived, and problems identified and discussed. Our analysis clearly shows that the PROSAIL model is well suited for the analysis of imaging spectrometer data from future satellite missions and that the model should be integrated in appropriate software tools that are being developed in this context for agricultural applications. The review supports the decision of potential users to employ PROSAIL for their specific data analysis and provides guidelines for choosing between the diverse retrieval techniques.
Land Use and Land Cover (LULC) classification is vital for environmental and ecological applications. Sentinel-2 is a new generation land monitoring satellite with the advantages of novel spectral capabilities, wide coverage and fine spatial and temporal resolutions. The effects of different spatial resolution unification schemes and methods on LULC classification have been scarcely investigated for Sentinel-2. This paper bridged this gap by comparing the differences between upscaling and downscaling as well as different downscaling algorithms from the point of view of LULC classification accuracy. The studied downscaling algorithms include nearest neighbor resampling and five popular pansharpening methods, namely, Gram-Schmidt (GS), nearest neighbor diffusion (NNDiffusion), PANSHARP algorithm proposed by Y. Zhang, wavelet transformation fusion (WTF) and high-pass filter fusion (HPF). Two spatial features, textural metrics derived from Grey-Level-Co-occurrence Matrix (GLCM) and extended attribute profiles (EAPs), are investigated to make up for the shortcoming of pixel-based spectral classification. Random forest (RF) is adopted as the classifier. The experiment was conducted in Xitiaoxi watershed, China. The results demonstrated that downscaling obviously outperforms upscaling in terms of classification accuracy. For downscaling, image sharpening has no obvious advantages than spatial interpolation. Different image sharpening algorithms have distinct effects. Two multiresolution analysis (MRA)-based methods, i.e., WTF and HFP, achieve the best performance. GS achieved a similar accuracy with NNDiffusion and PANSHARP. Compared to image sharpening, the introduction of spatial features, both GLCM and EAPs can greatly improve the classification accuracy for Sentinel-2 imagery. Their effects on overall accuracy are similar but differ significantly to specific classes. In general, using the spectral bands downscaled by nearest neighbor interpolation can meet the requirements of regional LULC applications, and the GLCM and EAPs spatial features can be used to obtain more precise classification maps.
The Sentinel-2A multi-spectral instrument (MSI) acquires multi-spectral reflective wavelength observations with directional effects due to surface reflectance anisotropy and changes in the solar and viewing geometry. Directional effects were examined by considering two ten day periods of Sentinel-2A data acquired close to the solar principal and orthogonal planes over approximately 20° × 10° of southern Africa. More than 6.6 million (January 2016) and 10.6 million (April 2016) pairs of reflectance observations sensed 3 or 7 days apart in the forward and backscatter directions in overlapping Sentinel-2A orbit swaths were considered. The Sentinel-2A data were projected into the MODIS sinusoidal projection but first had to be registered due to a misregistration issue evident in the overlapping orbits. The top of atmosphere reflectance data were corrected to surface reflectance using the SEN2COR atmospheric correction software. Only pairs of forward and backward reflectance values that were cloud and snow-free, unsaturated, and had no significant change in their 3 or 7 day separation, were considered. The maximum observed Sentinel-2A view zenith angle was 11.93°. Greater BRDF effects were apparent in the January data (acquired close to the solar principal plane) than the April data (acquired close to the orthogonal plane) and at higher view zenith angle. For the January data the average difference between the surface reflectance in the forward and backward scatter directions at the Sentinel-2A scan edges increased with wavelength from 0.035 (blue), 0.047 (green), 0.057 (red), 0.078 (NIR), to about 0.1 (SWIR). These differences may constitute a significant source of noise for certain applications.
Remote sensing has gained much attention for agronomic applications such as crop management or yield esti- mation. Crop phenotyping under field conditions has recently become another important application that re- quires specific needs: the considered remote-sensing method must be (1) as accurate as possible so that slight differences in phenotype can be detected and related to genotype, and (2) robust so that thousands of cultivars potentially quite different in terms of plant architecture can be characterized with a similar accuracy over different years and soil and weather conditions. In this study, the potential of nadir and off-nadir ground-based spectro-radiometric measurements to remotely sense five plant traits relevant for field phenotyping, namely, the leaf area index (LAI), leaf chlorophyll and nitrogen contents, and canopy chlorophyll and nitrogen contents, was evaluated over fourteen sugar beet (Beta vulgaris L.) cultivars, two years and three study sites. Among the di- versity of existing remote-sensing methods, two popular approaches based on various selected Vegetation Indices (VI) and PROSAIL inversion were compared, especially in the perspective of using them for phenotyping applications. Overall, both approaches are promising to remotely estimate LAI and canopy chlorophyll content (RMSE≤10%). In addition, VIs show a great potential to retrieve canopy nitrogen content (RMSE=10%). On the other hand, the estimation of leaf-level quantities is less accurate, the best accuracy being obtained for leaf chlorophyll content estimation based on VIs (RMSE=17%). As expected when observing the relationship be- tween leaf chlorophyll and nitrogen contents, poor correlations are found between VIs and mass-based or area- based leaf nitrogen content. Importantly, the estimation accuracy is strongly dependent on sun-sensor geometry, the structural and biochemical plant traits being generally better estimated based on nadir and off-nadir ob- servations, respectively. Ultimately, a preliminary comparison tends to indicate that, providing that enough samples are included in the calibration set, (1) VIs provide slightly more accurate performances than PROSAIL inversion, (2) VIs and PROSAIL inversion do not show significant differences in robustness across the different cultivars and years. Even if more data are still necessary to draw definitive conclusions, the results obtained with VIs are promising in the perspective of high-throughput phenotyping using UAV-embedded multispectral cameras, with which only a few wavebands are available.
Sentinel-2 is a wide-swath and fine spatial resolution satellite imaging mission designed for data continuity and enhancement of the Landsat and other missions. The Sentinel-2 data are freely available at the global scale, and have similar wavelengths and the same geographic coordinate system as the Landsat data, which provides an excellent opportunity to fuse these two types of satellite sensor data together. In this paper, a new approach is presented for the fusion of Landsat 8 Operational Land Imager and Sentinel-2 Multispectral Imager data to coordinate their spatial resolutions for continuous global monitoring. The 30 m spatial resolution Landsat 8 bands are downscaled to 10 m using available 10 m Sentinel-2 bands. To account for the land-cover/land-use (LCLU) changes that may have occurred between the Landsat 8 and Sentinel-2 images, the Landsat 8 panchromatic (PAN) band was also incorporated in the fusion process. The experimental results showed that the proposed approach is effective for fusing Landsat 8 with Sentinel-2 data, and the use of the PAN band can decrease the errors introduced by LCLU changes. By fusion of Landsat 8 and Sentinel-2 data, more frequent observations can be produced for continuous monitoring (this is particularly valuable for areas that can be covered easily by clouds, thereby, contaminating some Landsat or Sentinel-2 observations), and the observations are at a consistent fine spatial resolution of 10 m. The products have great potential for timely monitoring of rapid changes.
Leaf pigments provide valuable information about plant physiology. High resolution monitoring of their dynamics will give access to better understanding of processes occurring at different scales, and will be particularly important for ecologists, farmers, and decision makers to assess the influence of climate change on plant functions, and the adaptation of forest, crop, and other plant canopies. In this article, we present a new version of the widely-used PROSPECT model, hereafter named PROSPECT-D for dynamic, which adds anthocyanins to chlorophylls and carotenoids, the two plant pigments in the current version. We describe the evolution and improvements of PROSPECT-D compared to the previous versions, and perform a validation on various experimental datasets. Our results show that PROSPECT-D outperforms all the previous versions. Model prediction uncertainty is decreased and photosynthetic pigments are better retrieved. This is particularly the case for leaf carotenoids, the estimation of which is particularly challenging. PROSPECT-D is also able to simulate realistic leaf optical properties with minimal error in the visible domain, and similar performances to other versions in the near infrared and shortwave infrared domains.
Sentinel-2 is a very new programme of the European Space Agency (ESA) that is designed for fine spatial resolution global monitoring. Sentinel-2 images provide four 10 m bands and six 20 m bands. To provide more explicit spatial information, this paper aims to downscale the six 20 m bands to 10 m spatial resolution using the four directly observed 10 m bands. The outcome of this fusion task is the production of 10 Sentinel-2 bands with 10 m spatial resolution. This new fusion problem involves four fine spatial resolution bands, which is different to, and more complex than, the common pan-sharpening fusion problem which involves only one fine band. To address this, we extend the existing two main families of image fusion approaches (i.e., component substitution, CS, and multiresolution analysis, MRA) with two different schemes, a band synthesis scheme and a band selection scheme. Moreover, the recently developed area-to-point regression kriging (ATPRK) approach was also developed and applied for the Sentinel-2 fusion task. Using two Sentinel-2 datasets released online, the three types of approaches (eight CS and MRA-based approaches, and ATPRK) were compared comprehensively in terms of their accuracies to provide recommendations for the task of fusion of Sentinel-2 images. The downscaled ten-band 10 m Sentinel-2 datasets represent important and promising products for a wide range of applications in remote sensing. They also have potential for blending with the upcoming Sentinel-3 data for fine spatio-temporal resolution monitoring at the global scale.
Physically-based radiative transfer models (RTMs) help understand the interactions of radiation with vegetation and atmosphere. However, advanced RTMs can be computationally burdensome, which makes them impractical in many real applications, especially when many state conditions and model couplings need to be studied. To overcome this problem, it is proposed to substitute RTMs through surrogate meta-models also named emulators. Emulators approximate the functioning of RTMs through statistical learning regression methods, and can open many new applications because of their computational efficiency and outstanding accuracy. Emulators allow fast global sensitivity analysis (GSA) studies on advanced, computationally expensive RTMs. As a proof-of-concept, three machine learning regression algorithms (MLRAs) were tested to function as emulators for the leaf RTM PROSPECT-4, the canopy RTM PROSAIL, and the computationally expensive atmospheric RTM MODTRAN5. Selected MLRAs were: kernel ridge regression (KRR), neural networks (NN) and Gaussian processes regression (GPR). For each RTM, 500 simulations were generated for training and validation. The majority of MLRAs were excellently validated to function as emulators with relative errors well below 0.2%. The emulators were then put into a GSA scheme and compared against GSA results as generated by original PROSPECT-4 and PROSAIL runs. NN and GPR emulators delivered identical GSA results, while processing speed compared to the original RTMs doubled for PROSPECT-4 and tripled for PROSAIL. Having the emulator-GSA concept successfully tested, for six MODTRAN5 atmospheric transfer functions (outputs), i.e., direct and diffuse at-surface solar irradiance ( E d i f , E d i r ), direct and diffuse upward transmittance ( T d i r , T d i f ), spherical albedo (S) and path radiance ( L 0 ), the most accurate MLRA’s were subsequently applied as emulator into the GSA scheme. The sensitivity analysis along the 400–2500 nm spectral range took no more than a few minutes on a contemporary computer—in comparison, the same analysis in the original MODTRAN5 would have taken over a month. Key atmospheric drivers were identified, which are on the one hand aerosol optical properties, i.e., aerosol optical thickness (AOT), Angstrom coefficient (AMS) and scattering asymmetry variable (G), mostly driving diffuse atmospheric components, E d i f and T d i f ; and those affected by atmospheric scattering, L 0 and S. On the other hand, as expected, AOT, AMS and columnar water vapor (CWV) in the absorption regions mostly drive E d i r and T d i r atmospheric functions. The presented emulation schemes showed very promising results in replacing costly RTMs, and we think they can contribute to the adoption of machine learning techniques in remote sensing and environmental applications.
The major constraint in understanding grass above ground biomass variations using remotely sensed data are the expenses associated with the data, as well as the limited number of techniques that can be applied to different management practices with minimal errors. New generation multispectral sensors such as Sentinel 2 Multispectral Imager (MSI) are promising for effective rangeland management due to their unique spectral bands and higher signal to noise ratio. This study resampled hyperspectral data to spectral resolutions of the newly launched Sentinel 2 MSI and the recently launched Landsat 8 OLI for comparison purposes. Using Sparse partial least squares regression, the resampled data was applied in estimating above ground biomass of grasses treated with different fertilizer combinations of ammonium sulfate, ammonium nitrate, phosphorus and lime as well as unfertilized experimental plots. Sentinel 2 MSI derived models satisfactorily performed (R2=0.81, RMSEP=1.07kg/m2, RMSEP_rel=14.97) in estimating grass above ground biomass across different fertilizer treatments relative to Landsat 8 OLI (Landsat 8 OLI: R2=0.76, RMSEP=1.15kg/m2, RMSEP_rel=16.04). In comparison, hyperspectral data derived models exhibited better grass above ground biomass estimation across complex fertilizer combinations (R2=0.92, RMSEP=0.69kg/m2, RMSEP_rel=9.61). Although Sentinel 2 MSI bands and indices better predicted above ground biomass compared with Landsat 8 OLI bands and indices, there were no significant differences (α=0.05) in the errors of prediction between the two new generational sensors across all fertilizer treatments. The findings of this study portrays Sentinel 2 MSI and Landsat 8 OLI as promising remotely sensed datasets for regional scale biomass estimation, particularly in resource scarce areas. © 2015 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS).
The scale mismatch between remote sensing observations and state variables simulated by crop growth models decreases the reliability of crop yield estimates. To overcome this problem, we implemented a two-step data-assimilation approach: first, we generated a time series of 30-m-resolution leaf area index (LAI) by combining Moderate Resolution Imaging Spectroradiometer (MODIS) data and three Landsat TM images with a Kalman filter algorithm (the synthetic KF LAI series); second, the time series were assimilated into the WOFOST crop growth model to generate an ensemble Kalman filter LAI time series (the EnKF-assimilated LAI series). The synthetic EnKF LAI series then drove the WOFOST model to simulate winter wheat yields at 1-km resolution for pixels with wheat fractions of at least 50%. The county-level aggregated yield estimates were compared with official statistical yields. The synthetic KF LAI time series produced a more realistic characterization of LAI phenological dynamics. Assimilation of the synthetic KF LAI series produced more accurate estimates of regional winter wheat yield (R 2 = 0.43; root-mean-square error (RMSE) = 439 kg ha −1) than three other approaches: WOFOST without assimilation (determination coefficient R 2 = 0.14; RMSE = 647 kg ha −1), assimilation of Landsat TM LAI (R 2 = 0.37; RMSE = 472 kg ha −1), and assimilation of S-G filtered MODIS LAI (R 2 = 0.49; RMSE = 1355 kg ha −1). Thus, assimilating the synthetic KF LAI series into the WOFOST model with the EnKF strategy provides a reliable and promising method for improving regional estimates of winter wheat yield.
Leaf area index (LAI) and evapotranspiration (ET) are two crucial biophysical variables related to crop growth and grain yield. This study presents a crop model–data assimilation framework to assimilate the 1-km moderate resolution imaging spectroradiometer (MODIS) LAI and ET products (MCD15A3 and MOD16A2, respectively) into the soil water atmosphere plant (SWAP) model to assess the potential for estimating winter wheat yield at field and regional scales. Since the 1-km MODIS products generally underestimate LAI or ET values in fragmented agricultural landscapes due to scale effects and intrapixel heterogeneity, we constructed a new cost function by comparing the generalized vector angle between the observed and modeled LAI and ET time series during the growing season. We selected three parameters (irrigation date, irrigation depth, and emergence date) as the reinitialized parameters to be optimized by minimizing the cost function using the shuffled complex evolution method—University of Arizona (SCE-UA) optimization algorithm, and then used the optimized parameters as inputs into the SWAP model for winter wheat yield estimation. We used four data-assimilation schemes to estimate winter wheat yield at field and regional scales. We found that jointly assimilating MODIS LAI and ET data improved accuracy (${bf R}^{bf 2} = 0.43$, ${bf RMSE} = {619};{kg},{cdot} {bf ha}^{- 1}$) than assimilating MODIS LAI data (${bf R}^2 = 0.28$, ${bf RMSE} = {889};{bf kg};{cdot};{bf ha}^{- 1}$) or ET data (${bf R}^{2} = 0.36$, ${bf RMSE} = {bf 1561};{bf kg};{- dot};{bf ha}^{- 1}$) at the county level, which indicates that the proposed estimation method is reliable and applicable at a county scale.
Sen2Core is a prototype processor for Sentinel-2 Level 2A product processing and formatting. The processor is developed for and with ESA and performs the tasks of Atmospheric Correction and Scene Classification of Level 1C input data. Level 2A outputs are: Bottom-Of-Atmosphere (BOA) corrected reflectance images, Aerosol Optical Thickness-, Water Vapour-, Scene Classification maps and Quality indicators, including cloud and snow probabilities. The Level 2A Product Formatting performed by the processor follows the specification of the Level 1C User Product.
Optical remote sensing provides information on important vegetation variables such as leaf area index (LAI), biomass, and chlorophyll content. In this study, rice crops, which are rarely studied, were selected because of their high economic importance and the role they play in food security in the study area. The aim was to obtain a reliable estimate of canopy chlorophyll content as an important indicator for the evaluation of the plant status. PROSAIL radiative transfer model and the multispectral image data of ALOS AVNIR-2 were used. A field campaign was carried out in July 2010 in the northern part of Iran, Amol. Sixty sample plots of 20 × 20 m-2 were randomly selected, and their chlorophyll content was measured using a SPAD-502 chlorophyll meter. The PROSAIL was inverted using a lookup-table (LUT) approach. The LUTs were generated in different sizes. The effect of the LUT size on the retrieval accuracy of the canopy's chlorophyll content was studied using analysis of variance (ANOVA). The outcome of the inversion was evaluated using the calculated R2 and RMSE values with the field measurements. The obtained results demonstrate the ability of PROSAIL to estimate rice plant chlorophyll content using ALOS AVNIR-2 multispectral data (R2= 0.65; RMSE = 0.45). The results also confirmed the usefulness of such an approach for crop monitoring and ecological applications.
The aim of this work was to investigate different approaches for the estimation of canopy structure properties from multiangular measurements at the field scale. Hyperspectral multiangular data were acquired on potato canopies using a spectroradiometer (GER-1500) and corresponding multiangular images using the VIFIS (Variable Interference Filter Imaging Spectrometer). The data obtained using the spectroradiometer were employed in the inversion of the PROSAIL model. The images obtained from the VIFIS were classified into the component image fractions: shaded and sunlit leaves and soil. These classification results were then used directly in the inversion of a simple ray-tracing canopy model. The inversion technique was based on a look-up table approach using a simple ray-tracing model of a plant canopy. Field sampling was carried out for the direct measurement of leaf area index (LAI) and other canopy properties. The experimental error in the data of both sensors was large since the canopy appeared non-homogeneous at the measurement height used, mainly because of the crop row structure. However LAI values retrieved from both approaches were realistic and allowed the discrimination of potato canopies that had received different nitrogen fertilization treatments. The relative merits and practicalities of the two approaches (multiangular hyperspectral reflectance versus image classification) are discussed.
Sentinel-2 is planned for launch in 2014 by the European Space Agency
and it is equipped with the Multi Spectral Instrument (MSI), which will
provide images with high spatial, spectral and temporal resolution. It
covers the VNIR/SWIR spectral region in 13 bands and incorporates two
new spectral bands in the red-edge region, which can be used to derive
vegetation indices using red-edge bands in their formulation. These are
particularly suitable for estimating canopy chlorophyll and nitrogen (N)
content. This band setting is important for vegetation studies and is
very similar to the ones of the Ocean and Land Colour Instrument (OLCI)
on the planned Sentinel-3 satellite and the Medium Resolution Imaging
Spectrometer (MERIS) on Envisat, which operated from 2002 to early 2012.
This paper focuses on the potential of Sentinel-2 and Sentinel-3 in
estimating total crop and grass chlorophyll and N content by studying in
situ crop variables and spectroradiometer measurements obtained for four
different test sites. In particular, the red-edge chlorophyll index
(CIred-edge), the green chlorophyll index
(CIgreen) and the MERIS terrestrial chlorophyll index (MTCI)
were found to be accurate and linear estimators of canopy chlorophyll
and N content and the Sentinel-2 and -3 bands are well positioned for
deriving these indices. Results confirm the importance of the red-edge
bands on particularly Sentinel-2 for agricultural applications, because
of the combination with its high spatial resolution of 20 m.
In the framework of the Global Monitoring for Environment and Security (GMES) programme, the European Space Agency (ESA) in partnership with the European Union (EU) is developing the Sentinel-2 optical imaging mission devoted to the operational monitoring of land and coastal areas. The Sentinel-2 mission is based on a twin satellites configuration deployed in polar sun-synchronous orbit and designed to offer a unique combination of systematic global coverage, high revisit (five days at equator with two satellites) and high spatial resolution imagery (10/20/60m). The Multispectral instrument features 13 spectral bands, going from visible to short wave infrared domains. The instrument is designed to provide in orbit calibration, excellent radiometric and geometric performance, and with a capability to support accurate image geolocation and co-registration. The Sentinel-2 mission is more particularly tailored to the monitoring of land terrains, including vegetation and urban areas. Sentinel-2 will ensure data continuity with the SPOT and Landsat multi-spectral sensors, while accounting for future service evolution.
The robust and accurate retrieval of vegetation biophysical variables using radiative transfer models (RTM) is seriously hampered by the ill-posedness of the inverse problem. With this research we further develop our previously published (object-based) inversion approach [Atzberger, 2004, RSE 93: 53–67] and evaluate it against simulated Sentinel-2 data. The proposed RTM inversion takes advantage of the geostatistical fact that the biophysical characteristics of nearby pixels are generally more similar than those at a larger distance. This leads to spectral co-variations in the n-dimensional spectral features space, which can be used for regularization purposes. A simple two-step inversion based on PROSPECT + SAIL generated look-up-tables is presented that can be easily adapted to other radiative transfer models. The approach takes into account the spectral signatures of adjacent pixels in gliding (3 × 3) windows. Using a set of leaf area index (LAI) measurements (n = 26) acquired over alfalfa, sugar beet and garlic crops of the Barrax test site (Spain), it is demonstrated that the proposed regularization using neighbourhood information yields more accurate results compared to the pixel-based inversion. With the proposed regularization, the RMSE between ground measured and Sentinel-2 derived LAI is 0.54 m2/m2 and hence significantly lower compared to the RMSE of the standard inversion approach (RMSE: 1.46 m2/m2) and also of higher accuracy compared to a scaled NDVI model (RMSE: 0.90 m2/m2). At the same time, a positive effect on the modelled leaf chlorophyll content (Cab) is noticed, albeit too few field measurements were available for deriving statistically sound results. For the other retrieved biophysical parameters such as leaf dry matter content (Cm), soil brightness (αsoil) and average leaf angle (ALA) the proposed algorithm yields more plausible and spatially consistent results. Altogether the findings confirm the positive effect of regularizing the model inversion using spatial constraints. As for any other inversion strategy, the approach requires a RTM well suited for the crop under study. For three additional crops (maize, potatoes and sunflower), the forward modelling with field measured LAI did not match the observed signatures. Consequently, for these canopies both the standard and the object-based inversion failed.
The DART (discrete anisotropic radiative transfer) model simulates radiative transfer in heterogeneous 3-D scenes; here, a forest plantation. Similarly to Kimes model, the scene is divided into a rectangular cell matrix, i.e., a building block for simulating larger scenes. Cells are parallelipipedic. The scene encompasses different landscape features (i.e., trees with leaves and trunks, grass, water, and soil) with specific optical (reflectance, transmittance) and structural (LAI, LAD) characteristics. Radiation directions are subdivided into contiguous sectors with possibly uneven spacing. Topography, hot spot, and multiple interactions (scattering, attenuation) within cells are modeled. Two major steps are distinguished: (1) Illumination of cells by direct sun radiation. Actual locations of within cell scattering are determined for optimizing scattering computation. (2) Interception and scattering of previously scattered radiation. Diffuse atmospheric radiation is input at this level. Multiple scattering is represented with a spherical harmonic decomposition, for reducing data volume. The model iterates on step 2 for all cells, and stops with the energetic equilibrium. This model predicts the bi-directional reflectance factors of 3D canopies, with each scene component contribution; it was successfully tested with homogeneous covers. It gives also the radiation regime with canopies, and consequently some information about volume distribution of photosynthesis rates and primary production.
Statistical and physical models have seldom been compared in studying grasslands. In this paper, both modeling approaches are investigated for mapping leaf area index (LAI) in a Mediterranean grassland (Majella National Park, Italy) using HyMap airborne hyperspectral images. We compared inversion of the PROSAIL radiative transfer model with narrow band vegetation indices (NDVI-like and SAVI2-like) and partial least squares regression (PLS). To assess the performance of the investigated models, the normalized RMSE (nRMSE) and R2 between in situ measurements of leaf area index and estimated parameter values are reported. The results of the study demonstrate that LAI can be estimated through PROSAIL inversion with accuracies comparable to those of statistical approaches (R2 = 0.89, nRMSE = 0.22). The accuracy of the radiative transfer model inversion was further increased by using only a spectral subset of the data (R2 = 0.91, nRMSE = 0.18). For the feature selection wavebands not well simulated by PROSAIL were sequentially discarded until all bands fulfilled the imposed accuracy requirements.
The relationship between SPAD-501 meter readings (SPAD) and total chlorophyll content (TCHL) was evaluated for `Delicious' apple (Malus domestica Borkh.) leaves grown in various environments. Regression models were developed between SPAD and TCHL for each of six separate experiments and were evaluated for statistical coincidence. SPAD was linearly related in a positive manner to TCHL in five of the six experiments; however, models differed between experiments, particularly between field- and greenhouse-grown trees. Thus, the relationship between SPAD and TCHL must be determined for each experiment.
We use a simple radiative transfer model with vegetation, soil, and atmospheric components to illustrate how the normalized difference vegetation index (NDVI), leaf area index (LAI), and fractional vegetation cover are dependent. In particular, we suggest that LAI and fractional vegetation cover may not be independent quantitites, at least when the former is defined without regard to the presence of bare patches between plants, and that the customary variation of LAI with NDVI can be explained as resulting from a variation in fractional vegetation cover. The following points are made: i) Fractional vegetation cover and LAI are not entirely independent quantities, depending on how LAI is defined. Care must be taken in using LAI and fractional vegetation cover independently in a model because the former may partially take account of the latter; ii) A scaled NDVI taken between the limits of minimum (bare soil) and miximum fractional vegetation cover is insenstive to atmospheric correction for both clear and hazy conditions, at least for viewing angles less than about 20 degrees from nadir; iii) A simple relation between scaled NDVI and fractional vegetation cover, previously described in the literature, is further confirmed by the .simulations; iv) The sensitive dependence of LAI on NDVI when the former is below a value of about 2–4 may be viewed as being due to the variation in the bare soil component.
Farmers must balance the competing goals of supplying adequate N for their crops while minimizing N losses to the environment. To characterize the spatial variability of N over large fields, traditional methods (soil testing, plant tissue analysis, and chlorophyll meters) require many point samples. Because of the close link between leaf chlorophyll and leaf N concentration, remote sensing techniques have the potential to evaluate the N variability over large fields quickly. Our objectives were to (1) select wavelengths sensitive to leaf chlorophyll concentration, (2) simulate canopy reflectance using a radiative transfer model, and (3) propose a strategy for detecting leaf chlorophyll status of plants using remotely sensed data. A wide range of leaf chlorophyll levels was established in field-grown corn (Zea mays L.) with the application of 8 N levels: 0%, 12.5%, 25%, 50%, 75%, 100%, 125%, and 150% of the recommended rate. Reflectance and transmittance spectra of fully expanded upper leaves were acquired over the 400-nm to 1,000-nm wavelength range shortly after anthesis with a spectroradiometer and integrating sphere. Broad-band differences in leaf spectra were observed near 550 nm, 715 nm, and >750 nm. Crop canopy reflectance was simulated using the SAIL (Scattering by Arbitrarily Inclined Leaves) canopy reflectance model for a wide range of background reflectances, leaf area indices (LAI), and leaf chlorophyll concentrations. Variations in background reflectance and LAI confounded the detection of the relatively subtle differences in canopy reflectance due to changes in leaf chlorophyll concentration. Spectral vegetation indices that combined near-infrared reflectance and red reflectance (e.g., OSAVI and NIR/Red) minimized contributions of background reflectance, while spectral vegetation indices that combined reflectances of near-infrared and other visible bands (MCARI and NIR/Green) were responsive to both leaf chlorophyll concentrations and background reflectance. Pairs of these spectral vegetation indices plotted together produced isolines of leaf chlorophyll concentrations. The slopes of these isolines were linearly related to leaf chlorophyll concentration. A limited test with measured canopy reflectance and leaf chlorophyll data confirmed these results. The characterization of leaf chlorophyll concentrations at the field scale without the confounding problem of background reflectance and LAI variability holds promise as a valuable aid for decision making in managing N applications.
To estimate regional-scale winter wheat (Triticum aestivum) yield, we developed a data-assimilation scheme that
assimilates remotely sensed reflectance into a coupled crop growth–radiative transfer model. We generated a
time series of 8-day, 30-m-resolution synthetic Kalman Smoothed reflectance by combining MODIS surface reflectance
products with Landsat surface reflectance using a KS algorithm. We evaluated the assimilation performance
using datasets with different spatial and temporal scales (e.g., three dates for the 30-m Landsat reflectance,
8-day and 1-km MODIS surface reflectance, and 8-day and 30-m synthetic KS reflectance) into the
coupled WOFOST–PROSAIL model. Then we constructed a four-dimensional variational data assimilation
(4DVar) cost function to account for differences between the observed and simulated reflectance. We used the
shuffled complex evolution–University of Arizona (SCE-UA) algorithm to minimize the 4DVar cost function and
optimize important input parameters of the coupled model. The optimized parameters were used to drive
WOFOST and estimate county-level winter wheat yield in a region of China. By assimilating the synthetic KS
reflectance data, we achieved the most accurate yield estimates (R2=0.44, 0.39, and 0.30; RMSE=598, 1288,
and 595 kg/ha for 2009, 2013, and 2014, respectively), followed by Landsat reflectance (R2=0.21, 0.22, and
0.33; RMSE=915, 1422, and 637 kg/ha for 2009, 2013, and 2014, respectively) and MODIS reflectance
(R2=0.49, 0.05, and 0.22; RMSE=1136, 1468, and 700 kg/ha for 2009, 2013, and 2014, respectively) at the
county level. Thus, our method improves the reliability of regional-scale crop yield estimates.
Large-scale crop monitoring and yield estimation are important for both scientific research and practical applications. Satellite remote sensing provides an effective means for regional and global cropland monitoring, particularly in data-sparse regions that lack reliable ground observations and reporting. The conventional approach of using visible and near-infrared based vegetation index (VI) observations has prevailed for decades since the onset of the global satellite era. However, other satellite data encompass diverse spectral ranges that may contain complementary information on crop growth and yield, but have been largely understudied and underused. Here we conducted one of the first attempts at synergizing multiple satellite data spanning a diverse spectral range, including visible, near-infrared, thermal and microwave, into one framework to estimate crop yield for the U.S. Corn Belt, one of the world's most important food baskets. Specifically, we included MODIS Enhanced VI (EVI), estimated Gross Primary Production based on GOME-2 solar-induced fluorescence (SIF-GPP), thermal-based ALEXI Evapotranspiration (ET), QuikSCAT Ku-band radar backscatter, and AMSR-E X-band passive microwave Vegetation Optical Depth (VOD) in this study, benchmarked on USDA county-level crop yield statistics. We used Partial Least Square Regression (PLSR), an effective statistical model for dimension reduction, to distinguish commonly shared and unique individual information from the various satellite data and other ancillary climate information for crop yield estimation. In the PLSR model that includes all of the satellite data and climate variables from 2007 to 2009, we assessed the first two major PLSR components and found that the first component (an integrated proxy of crop aboveground biomass) explained 82% variability of modelled crop yield, and the second component (dominated by environmental stresses) explained 15% variability of modelled crop yield. We found that most of the satellite derived metrics (e.g. SIF-GPP, radar backscatter, EVI, VOD, ALEXI-ET) share common information related to aboveground crop biomass (i.e. the first component). For this shared information, the SIF-GPP and backscatter data contain almost the same amount of information as EVI at the county scale. When removing the above shared component from all of the satellite data, we found that EVI and SIF-GPP do not provide much extra information; instead, Ku-band backscatter, thermal-based ALEXI-ET, and X-band VOD provide unique information on environmental stresses that improves overall crop yield predictive skill. In particular, Ku-band backscatter and associated differences between morning and afternoon overpasses contribute unique information on crop growth and environmental stress. Overall, using satellite data from various spectral bands significantly improves regional crop yield predictions. The additional use of ancillary climate data (e.g. precipitation and temperature) further improves model skill, in part because the crop reproductive stage related to harvest index is highly sensitive to environmental stresses but they are not fully captured by the satellite data used in our study. We conclude that using satellite data across various spectral ranges can improve monitoring of large-scale crop growth and yield beyond what can be achieved from individual sensors. These results also inform the synergistic use and development of current and next generation satellite missions, including NASA ECOSTRESS, SMAP, and OCO-2, for agricultural applications.
Real-time, nondestructive monitoring of crop nitrogen (N) status is important for precise N management in winter wheat production. Nadir viewing passive multispectral sensors have limited utility for measuring the N status of winter wheat in middle and bottom layers, and multi-angular remote sensors may instead improve detection of whole canopy physiological and biochemical parameters. Our objective was to improve the predictive accuracy and angular stability of leaf nitrogen concentration (LNC) measurement by constructing a novel Angular Insensitivity Vegetation Index (AIVI). We quantified the relationship between LNC and ground-based multi-angular hyperspectral reflectance in winter wheat (Triticum aestivum L.) across different growth stages, plant types, N rates, planting density, ecological sites and years. The optimum vegetation indices (VIs) obtained from 17 traditional indices reported in the literature were tested for their stability in estimating LNC at 13 view zenith angles (VZAs) in the solar principal plane (SPP). Overall the back-scatter direction gave improved index performance, relative to the nadir and forward-scattering direction. Red-edge VIs (e.g., mND705, GND [750,550], NDRE, RI-1dB) were highly correlated with LNC. However, the relationships strongly depended on experimental conditions, and these VIs tended to saturate at the highest LNC (4.5%). To further overcome the influence of different experimental conditions and VZAs on VIs, we developed a novel index, Angular Insensitivity Vegetation Index (AIVI), based on red-edge, blue and green bands. Our new model showed the highest association with LNC (R2=0.73-0.87) compared to traditional VIs. Investigating AIVI predictive accuracy in measuring LNC across view zenith angles (VZAs) revealed that performance was the highest at -20° and was relatively homogenous between -10° and -40°. This provided a united, predictive model across this wide-angle range, which enhances the possibility of N monitoring by using portable monitors. Testing of the models with independent data gave R2 of 0.84 at -20°, and 0.83 across the range of -10° to -40°, respectively. These results suggest that the novel AIVI is more effective for monitoring LNC than previously reported VIs for predicting accuracy, monitoring model stability and view angle independency. More generally, our model indicates the importance of accounting for angular effects when analyzing VIs under different experimental conditions.
Use of leaf meters to provide an instantaneous assessment of leaf chlorophyll has become common, but calibration of meter output into direct units of leaf chlorophyll concentration has been difficult and an understanding of the relationship between these two parameters has remained elusive. We examined the correlation of soybean (Glycine max) and maize (Zea mays L.) leaf chlorophyll concentration, as measured by organic extraction and spectrophotometric analysis, with output (M) of the Minolta SPAD-502 leaf chlorophyll meter. The relationship is non-linear and can be described by the equation chlorophyll (μmol m−2)=10(M0.265), r
2=0.94. Use of such an exponential equation is theoretically justified and forces a more appropriate fit to a limited data set than polynomial equations. The exact relationship will vary from meter to meter, but will be similar and can be readily determined by empirical methods. The ability to rapidly determine leaf chlorophyll concentrations by use of the calibration method reported herein should be useful in studies on photosynthesis and crop physiology.
Accurate measurement of leaf area index (LAI), an important characteristic of plant canopies directly linked to primary production, is essential for monitoring changes in ecosystem C stocks and other ecosystem level fluxes. Direct measurement of LAI is labor intensive, impractical at large scales and does not capture seasonal or annual variations in canopy biomass. The need to monitor canopy related fluxes across landscapes makes remote sensing an attractive technique for estimating LAI. Many vegetation indices, such as Normalized Difference Vegetation Index (NDVI), tend to saturate at LAI levels > 4 although tropical and temperate forested ecosystems often exceed that threshold. Using two monospecific shrub thickets as model systems, we evaluated the potential of a variety of algorithms specifically developed to improve accuracy of LAI estimates in canopies where LAI exceeds saturation levels for other indices. We also tested the potential of indices developed to detect variations in canopy chlorophyll to estimate LAI because of the direct relationship between total canopy chlorophyll content and LAI. Indices were evaluated based on data from direct (litterfall) and indirect measurements (LAI-2000) of LAI. Relationships between results of direct and indirect ground-sampling techniques were also evaluated. For these two canopies, the indices that showed the highest potential to accurately differentiate LAI values > 4 were derivative indices based on red-edge spectral reflectance. Algorithms intended to improve accuracy at high LAI values in agricultural systems were insensitive when LAI exceeded 4 and offered little or no improvement over NDVI. Furthermore, indirect ground-sampling techniques often used to evaluate the potential of vegetation indices also saturate when LAI exceeds 4. Comparisons between hyperspectral vegetation indices and a saturated LAI value from indirect measurement may overestimate accuracy and sensitivity of some vegetation indices in high LAI communities. We recommend verification of indirect measurements of LAI with direct destructive sampling or litterfall collection, particularly in canopies with high LAI.
\textless}p{\textgreater}{\textless}br/{\textgreater}The} aim of this paper was to extend the method of downscaling cokriging for image fusion by making the method spatially adaptive in that the filter parameters (cokriging weights) can change across the image. The method can adapt itself to the usual statistical non-homogeneity (spatially variable mean, variance and correlation length) of a satellite sensor image that covers an area with different spatial patterns of geographical objects or different terrain types. The solution adopted was to estimate the models of covariances and cross-covariances (or semivariograms and cross-semivariograms) by the same procedure as described in {Pardo-Iguzquiza} et al. (2006) but with the method applied locally instead of globally. The correct implementation of this local estimation is the key for computational feasibility and prediction efficiency. Two parameters to be taken into account are the grid of locations on which a moving window is centred (local modelling is performed inside this window) and the size of this moving window. With respect to the latter parameter, there is a trade-off between a size small enough to make the procedure locally adaptive and large enough to produce reliable statistical estimates. The computational burden will impose limits to the distance between grid points on which the local moving window is centred. A case study with Landsat {ETM+} images was used to show the implementation of the method and the result was evaluated using several statistics widely used for assessing the quality of a fused image, apart from its visual appearance.{\textless}/p{\textgreater
Leaf area index (LAI) and leaf angle distribution are widely used indices of vegetative canopy structure that are difficult to measure directly. This study was conducted to test a commercially available instrument for rapidly determining LAI and foliage inclination information from “fisheye” measurements of light interception. The instrument's estimates of LAI are compared with direct measurements in soybean [Glycine max (L.) Merr.], winter wheat (Triticum aestivum L.), and prairie grass. The dominant grass species in the plots were Indian grass [Sorghastrum nutans (L.) Nash], switchgrass (Panicum virgatum L.), and big bluestem (Andropogon gerardii Vitman). The instrument's LAI resolution was better than 3%, and its LAI error was generally less than 15%. Variations in sky brightness patterns caused variations in LAI estimates in winter wheat of less than 10%, and the presence of direct solar radiation increased LAI errors to more than 30% in canopies of differing species and LAI. In the presence of gaps in the canopy, the sensor's azimuthal view should be reduced. A simple test indicates if a canopy's gaps are significant.
Please view the pdf by using the Full Text (PDF) link under 'View' to the left. Copyright © . .
Remotely sensed vegetation indices such as NDVI, computed using the red and near infrared bands have been used to estimate pasture biomass. These indices are of limited value since they saturate in dense vegetation. In this study, we evaluated the potential of narrow band vegetation indices for characterizing the biomass of Cenchrus ciliaris grass measured at high canopy density. Three indices were tested: Modified Normalized Difference Vegetation Index (MNDVI), Simple Ratio (SR) and Transformed Vegetation Index (TVI) involving all possible two band combinations between 350 nm and 2500 nm. In addition, we evaluated the potential of the red edge position in estimating biomass at full canopy cover. Results indicated that the standard NDVI involving a strong chlorophyll absorption band in the red region and a near infrared band performed poorly in estimating biomass (R2=0.26). The MNDVIs involving a combination of narrow bands in the shorter wavelengths of the red edge (700-750 nm) and longer wavelengths of the red edge (750-780 nm), yielded higher correlations with biomass (mean R2=0.77 for the highest 20 narrow band NDVIs). When the three vegetation indices were compared, SR yielded the highest correlation coefficients with biomass as compared to narrow band NDVI and TVI (average R2=0.80, 0.77 and 0.77 for the first 20 ranked SR, NDVI and TVI respectively). The red edge position yielded comparable results to the narrow band vegetation indices involving the red edge bands. These results indicate that at high canopy density, pasture biomass may be more accurately estimated by vegetation indices based on wavelengths located in the red edge than the standard NDVI.
Mapping and monitoring of leaf area index (LAI) is important for spatially distributed modeling of vegetation productivity, evapotranspiration, and surface energy balance. Global LAI surfaces will be an early product of the MODIS Land Science Team, and the requirements for LAI validation at selected sites have prompted interest in accurate LAI mapping at a more local scale. While spectral vegetation indices (SVIs) derived from satellite remote sensing have been used to map LAI, vegetation type, and related optical properties, and effects of Sun–surface–sensor geometry, background reflectance, and atmospheric quality can limit the strength and generality of empirical LAI–SVI relationships. In the interest of a preliminary assessment of the variability in LAI–SVI relationships across vegetation types, we compared Landsat 5 Thematic Mapper imagery from three temperate zone sites with on-site LAI measurements. The sites differed widely in location, vegetation physiognomy (grass, shrubs, hardwood forest, and conifer forest), and topographic complexity. Comparisons were made using three different red and near-infrared-based SVIs (NDVI, SR, SAVI). Several derivations of the SVIs were examined, including those based on raw digital numbers (DN), radiance, top of the atmosphere reflectance, and atmospherically corrected reflectance. For one of the sites, which had extreme topographic complexity, additional corrections were made for Sun–surface–sensor geometry. Across all sites, a strong general relationship was preserved, with SVIs increasing up to LAI values of 3 to 5. For all but the coniferous forest site, sensitivity of the SVIs was low at LAI values above 5. In coniferous forests, the SVIs decreased at the highest LAI values because of decreasing near-infrared reflectance associated with the complex canopy in these mature to old-growth stands. The cross-site LAI–SVI relationships based on atmospherically corrected imagery were stronger than those based on DN, radiance, or top of atmosphere reflectance. Topographic corrections at the conifer site altered the SVIs in some cases but had little effect on the LAI–SVI relationships. Significant effects of vegetation properties on SVIs, which were independent of LAI, were evident. The variability between and around the best fit LAI–SVI relationships for this dataset suggests that for local accuracy in development of LAI surfaces it will be desirable to stratify by land cover classes (e.g., physiognomic type and successional stage) and to vary the SVI.
Reflectance data in the green, red and near-infrared wavelength region were acquired by the SPOT high resolution visible and geometric imaging instruments for an agricultural area in Denmark (56°N, 9°E) for the purpose of estimating leaf chlorophyll content (Cab) and green leaf area index (LAI). SPOT reflectance observations were atmospherically corrected using aerosol data from MODIS and profiles of air temperature, humidity and ozone from the Atmospheric Infrared Sounder (AIRS), and used as input for the inversion of a canopy reflectance model. Computationally efficient inversion schemes were developed for the retrieval of soil and land cover-specific parameters which were used to build multiple species and site dependent formulations relating the two biophysical properties of interest to vegetation indices or single spectral band reflectances. Subsequently, the family of model generated relationships, each a function of soil background and canopy characteristics, was employed for a fast pixel-wise mapping of Cab and LAI.
The PROSPECT leaf optical model has, to date, combined the effects of photosynthetic pigments, but a finer discrimination among the key pigments is important for physiological and ecological applications of remote sensing. Here we present a new calibration and validation of PROSPECT that separates plant pigment contributions to the visible spectrum using several comprehensive datasets containing hundreds of leaves collected in a wide range of ecosystem types. These data include leaf biochemical (chlorophyll a, chlorophyll b, carotenoids, water, and dry matter) and optical properties (directional–hemispherical reflectance and transmittance measured from 400 nm to 2450 nm). We first provide distinct in vivo specific absorption coefficients for each biochemical constituent and determine an average refractive index of the leaf interior. Then we invert the model on independent datasets to check the prediction of the biochemical content of intact leaves. The main result of this study is that the new chlorophyll and carotenoid specific absorption coefficients agree well with available in vitro absorption spectra, and that the new refractive index displays interesting spectral features in the visible, in accordance with physical principles. Moreover, we improve the chlorophyll estimation (RMSE = 9 µg/cm2) and obtain very encouraging results with carotenoids (RMSE = 3 µg/cm2). Reconstruction of reflectance and transmittance in the 400–2450 nm wavelength domain using PROSPECT is also excellent, with small errors and low to negligible biases. Improvements are particularly noticeable for leaves with low pigment content.
Airborne multispectral data were acquired by the Compact Airborne Spectral Imager (CASI) for an agricultural area in Denmark with the purpose of quantifying vegetation amount and variations in the physiological status of the vegetation. Spectral reflectances, vegetation indices, and red edge positions were calculated on the basis of the CASI data and compared to field measurements of green leaf area index (LAI; L) and canopy nitrogen concentrations (Nc) sampled at 16 sites. Because of the variety of the samples with respect to vegetation type, leaf age, and phenological developmental stage, the data of L and Nc were uncorrelated. The scattering effect of leaves effectuated a robust linear relationship between L and near-infrared (NIR) reflectance (r=.93), whereas the Nc (vegetative period) was significantly correlated with the spectral reflectance in the green (r=−.88) and far-red wavebands (r=−.94). The correlations between vegetation indices and L were also important, in particular, for the enhanced vegetation index (EVI; r=.88), whereas the red edge position correlated less significantly with Nc (r=.78). Assuming L and Nc to be responsible for most of the spatial variability in the CO2 assimilation rates, remote sensing-based maps of these variables were produced for use in a coupled sun/shade photosynthesis/transpiration model. The predicted rates of net photosynthesis and transpiration compared reasonably with eddy covariance measurements of CO2 and water vapour fluxes recorded at four different crop fields. The results allowed evaluation of the spatial variations in the photosynthetic light, nitrogen, and water use efficiencies. While photosynthesis was linearly related to the transpiration, the light use efficiency (LUE) was found to be dependent on nitrogen concentrations.
Markov properties of stand geometry are incorporated into an analytical multispectral canopy reflectance model. The correlation of leaf positions in adjacent layers significantly influences the gap probability in a stand and, consequently, the canopy reflectance (CR) and its angular distribution. The sensitivity analysis demonstrated that the effect is greatest on the angular distribution of multiply scattered radiation. Validation of the new CR model that considers the Markov stand geometry demonstrated an improved agreement between model calculations and measured directional reflectance distribution of near infrared (NIR) reflectance for barley and soybean stands. As a consequence, inversion of the new model in the NIR spectral region and in two spectral bands simultaneously allowed significantly better estimation of the leaf area index of a stand than the Nilson-Kuusk model which assumes a Poisson stand geometry.
The potential of radiative transfer modelling and inversion techniques for operational uses is investigated in order to retrieve leaf area index in a poplar plantation. The 1-D bidirectional canopy reflectance model SAIL, coupled with the leaf optical properties model PROSPECT, was inverted with hyperspectral airborne DAIS data by means of an iterative method. The root mean square error in LAI estimation was determined against in situ measurements in order to evaluate the impact of different inversion strategies on the LAI retrieval accuracy. These included the selection of an optimal spectral sampling set, the exploitation of prior knowledge in the inversion process and the use of multiview angle data. We claim that the best configuration is achieved by exploiting multiview DAIS data and prior knowledge information about the model variables (RMSE of 0.39 m2 m−2). It is also shown that the use of prior knowledge and the selection of a limited number of bands forming the optimal spectral sampling are instrumental in increasing the accuracy of the inversion process. Our analysis confirms the operational potential of model inversion for biophysical parameter retrieval.
Remote sensing is a powerful tool for obtaining important agronomic information about field crops. Many spectral vegetation indices (VIs) have been developed in the past three decades to provide more sensitive measurements of plant biophysical parameters and to reduce external noise interferences such as those related to soil and the atmosphere. Some VIs were developed based on narrowband spectral data and others on broadband sensors. Therefore, although the mathematical equations defining VIs are the same, their calculated values are different, thus affecting their stability in predicting agronomic variables such as total green leaf area index. The objective of this study was to compare the ability of VIs derived from broad and narrowbands and to determine the optimum red–NIR bands for VIs used in predicting leaf area index (LAI) and canopy chlorophyll density (CCD) of cotton canopies. A completely randomized experiment was conducted in a cotton (Gossypium hirsutum L. cv. Sumian 3) field treated with four nitrogen application rates: 0%, 50%, 100% and 200% of the recommended rate. Hyperspectral reflectance was measured at 2.3 m above the cotton canopy on July 15, August 14 and October 1, 2002 using a FieldSpec® FR spectroradiometer. Corresponding leaf area index values and CCD were also measured on these dates. A large number (i.e. 22,500) of two-band combinations in the Normalized Difference Vegetation Index (λ2 − λ1)/(λ1 + λ2) and the Ratio Vegetation index λ2/λ1 was used for a linear and exponential regression analysis against LAI and CCD values. Moreover, traditional broadband vegetation indices based on simulated spectra were compared with their narrowband versions in predicting LAI and CCD. The results suggest that 640–660 nm and 800–870 nm, the centers of the red and NIR channels of several multi-spectral sensors on the current generation of earth-orbiting satellites, were not always the optimum wavelength position of red–NIR bands for VIs. Although different in formula, both the NDVI (normalized difference vegetation index) and RVI (ratio vegetation index) calculated from narrowbands at 690–710 nm and 750–900 nm were closely correlated with LAI (R2 > 0.8) and CCD (R2 > 0.85). The red–NIR band position was more important than band width for modeling LAI and CCD. In summary, hyperspectral remotely sensed data provide more alternative red–NIR bands compared to multi-spectral data and, therefore, can provide greater flexibility in predicting LAI and CCD.
The Medium Resolution Imaging Spectrometer (MERIS), one of the payloads on Envisat, has fine spectral resolution, moderate spatial resolution and a 3-day repeat cycle. This makes MERIS a potentially valuable sensor for the measurement and monitoring of terrestrial environments at regional to global scales. The red edge, which results from an abrupt reflectance change in red and near-infrared (NIR) wavelengths has a location that is related directly to the chlorophyll content of vegetation. A new index called the MERIS terrestrial chlorophyll index (MTCI) uses data in three red/NIR wavebands centered at 681.25, 708.75 and 753.75 nm (bands 8, 9 and 10 in the MERIS standard band setting). The MTCI is easy to calculate and can be automated. Preliminary indirect evaluation using model, field and MERIS data suggested its sensitivity to chlorophyll content, notably at high values. As a result this index is now a standard level-2 product of the European Space Agency.
Accurate estimates of vegetation biophysical variables are valuable as input to models describing the exchange of carbon dioxide and energy between the land surface and the atmosphere and important for a wide range of applications related to vegetation monitoring, weather prediction, and climate change. The present study explores the benefits of combining vegetation index and physically based approaches for the spatial and temporal mapping of green leaf area index (LAI), total chlorophyll content (TCab), and total vegetation water content (VWC). A numerical optimization method was employed for the inversion of a canopy reflectance model using Terra and Aqua MODIS multi-spectral, multi-temporal, and multi-angle reflectance observations to aid the determination of vegetation-specific physiological and structural canopy parameters. Land cover and site-specific inversion modeling was applied to a restricted number of pixels to build multiple species- and environmentally dependent formulations relating the three biophysical properties of interest to a number of selected simpler spectral vegetation indices (VI). While inversions generally are computationally slow, the coupling with the simple and computationally efficient VI approach makes the combined retrieval scheme for LAI, TCab, and VWC suitable for large-scale mapping operations. In order to facilitate application of the canopy reflectance model to heterogeneous forested areas, a simple correction scheme was elaborated, which was found to improve forest LAI predictions significantly and also provided more realistic values of leaf chlorophyll contents.
An analytical reflectance model for a statistically homogeneous plant canopy has been developed. The most specific characteristics of the model are: 1) considering both the single and the multiple scattering of radiation in the canopy and on the soil and 2) accounting for the specular reflection of radiation on leaves and canopy hot spot. For the inversion of the model the technique suggested by Goel and Strebel (1983) has been applied. The reflectance model fits well the results of measurements both of the seasonal course of the nadir reflectance and of the angular distribution of the directional reflectance of the winter wheat and barley canopies.
In this paper, we present a theoretical and modeling framework to estimate the fractions of photosynthetically active radiation (PAR) absorbed by vegetation canopy (FAPARcanopy), leaf (FAPARleaf ), and chlorophyll (FAPARchl), respectively. FAPARcanopy is an important biophysical variable and has been used to estimate gross and net primary production. However, only PAR absorbed by chlorophyll is used for photosynthesis, and therefore there is a need to quantify FAPARchl. We modified and coupled a leaf radiative transfer model (PROSPECT) and a canopy radiative transfer model (SAIL-2), and incorporated a Markov Chain Monte Carlo (MCMC) method (the Metropolis algorithm) for model inversion, which provides probability distributions of the retrieved variables. Our two-step procedure is: (1) to retrieve biophysical and biochemical variables using coupled PROSPECT + SAIL-2 model (PROSAIL-2), combined with multiple daily images (five spectral bands) from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor; and (2) to calculate FAPARcanopy, FAPARleaf and FAPARchl with the estimated model variables from the first step. We evaluated our approach for a temperate forest area in the Northeastern US, using MODIS data from 2001 to 2003. The inverted PROSAIL-2 fit the observed MODIS reflectance data well for the five MODIS spectral bands. The estimated leaf area index (LAI) values are within the range of field measured data. Significant differences between FAPARcanopy and FAPARchl are found for this test case. Our study demonstrates the potential for using a model such as PROSAIL-2, combined with an inverse approach, for quantifying FAPARchl, FAPARleaf, FAPARcanopy, biophysical variables, and biochemical variables for deciduous broadleaf forests at leaf- and canopy-levels over time.
This paper aims to link the spectral and directional variations of the leaf Bidirectional Reflectance Distribution Function (BRDF) by differentiating specular and diffuse components. To do this, BRDF of laurel (Prunus laurocesarus), European beech (Fagus silvatica) and hazel (Corylus avellana) leaves were measured at 400 wavelengths evenly spaced over the visible (VIS) and near-infrared (NIR) domains (480–880 nm) and at 400 source-leaf-sensor configurations. Measurement analysis suggested a spectral invariance of the specular component, the directional shape of which was mainly driven by leaf surface roughness. A three-parameter physically based model was fitted on the BRDF at each wavelength, confirming the spectral invariance of the specular component in the VIS, followed by a slight deterioration in the NIR. Due to this component, the amount of reflected light which did not penetrate into the leaf, could be considered as significant at wavelengths of chlorophyll absorption. Finally, by introducing the PROSPECT model, we proposed a five-parameter model to simulate leaf spectral and bidirectional reflectance in the VIS–NIR.
This article describes the algorithmic principles used to generate LAI, fAPAR and fCover estimates from VEGETATION observations. These biophysical variables are produced globally at 10 days temporal sampling interval under lat–lon projection at 1/112° spatial resolution. After a brief description of the VEGETATION sensors, radiometric calibration process, based on vicarious desertic targets is first presented. The cloud screening algorithm was then fine tuned using a global network of cloudiness observations. Atmospheric correction is then achieved using the SMAC code with inputs coming from meteorological values of pressure, ozone and water vapour. Aerosol optical thickness is derived from MODIS climatology assuming continental aerosol type. The Roujean BRDF model is then adjusted for red, near infrared and short wave infrared bands used to the remaining cloud free observations collected over a time window of ± 15 days. Outliers due to possible cloud contamination or residual atmospheric correction are iteratively eliminated and prior information is used to get more robust estimates of the three BRDF kernel coefficients. Nadir viewing top of canopy reflectance in the three bands is input to the biophysical algorithm to compute the products at 10 days sampling interval. This algorithm is based on training neural networks over SAIL + PROPSPECT radiative transfer model simulations for each biophysical variable. Details on the way the training data base was generated and the neural network designed and calibrated are presented. Finally, theoretical performances are discussed. Validation over ground measurement data sets and inter-comparison with other similar biophysical products are presented and discussed in a companion paper. The CYCLOPES products and associated detailed documentation are available at http://postel.mediasfrance.org.
Estimation of canopy biophysical variables from remote sensing data was investigated using radiative transfer model inversion. Measurement and model uncertainties make the inverse problem ill posed, inducing difficulties and inaccuracies in the search for the solution. This study focuses on the use of prior information to reduce the uncertainties associated to the estimation of canopy biophysical variables in the radiative transfer model inversion process. For this purpose, lookup table (LUT), quasi-Newton algorithm (QNT), and neural network (NNT) inversion techniques were adapted to account for prior information. Results were evaluated over simulated reflectance data sets that allow a detailed analysis of the effect of measurement and model uncertainties. Results demonstrate that the use of prior information significantly improves canopy biophysical variables estimation. LUT and QNT are sensitive to model uncertainties. Conversely, NNT techniques are generally less accurate. However, in our conditions, its accuracy is little dependent significantly on modeling or measurement error. We also observed that bias in the reflectance measurements due to miscalibration did not impact very much the accuracy of biophysical estimation.
A neural network is developed to operationally estimate biophysical variables over land surfaces from the observations of the ENVISAT-MERIS instrument: the leaf area index (LAI), the fraction of absorbed photosynthetically active radiation (fAPAR), the fraction of vegetation cover (fCover), and the canopy chlorophyll content (LAI×Cab). The neural network requires as input the geometry of observation and the top of canopy reflectances, corrected from the atmospheric effects, in eleven spectral bands. It is trained on a reflectance database made of radiative transfer model simulations. The principles underlying the generation of the database and the design of the network are first presented. The estimated variables are then compared to other existing products, LAI- and fAPAR-MODIS and MGVI-MERIS, and validated against ground measurements performed in the framework of the VALERI project. Results show remarkable consistency of the temporal dynamics between the several products with however some differences in the range of variation. When compared to actual VALERI ground measurements, the proposed algorithm shows the best performances for LAI (RMSE = 0.47) and fAPAR (RMSE = 0.09).