Resumen: Sentinel-2 (S2) es un nuevo satélite de la ESA que cuenta con 13 bandas proporcionando imágenes de alta calidad radiométrica y excelente resolución espacial (10 y 20 m) ideal para trabajos de clasificación. En este trabajo se han abordado dos objetivos: determinar el mejor método de clasificación con S2, y cuantificar su mejora respecto a otras misiones operativas, como SPOT. Para ello se han seleccionado cuatro clasificadores (LDA, RF, Árboles de decisión, K-NN) que se han aplicado en dos zonas agrarias: una en la huerta de Valencia (España) y otra en la región de Buenos Aires (Argentina). Se han probado todos los clasificadores usando, por una parte, todas las bandas de S2, y por otra usando sólo las cuatro que coinciden con SPOT. En todos los casos se han aplicando porcentajes entre el 10 y el 50% de datos de entrenamiento y usado el resto de datos como validación. Como resultado se ha generado un mapa de usos del suelo a partir del mejor clasificador, basándose en el índice Kappa, proporcionando información científicamente relevante como es el área ocupada por cada una de las clases.Palabras clave: clasificación, teledetección, usos del suelo, índice Kappa, Sentinel-2. Land use classification from Sentinel-2 imageryAbstract: Sentinel-2 (S2), a new ESA satellite for Earth observation, accounts with 13 bands which provide high-quality radiometric images with an excellent spatial resolution (10 and 20 m) ideal for classification purposes. In this paper, two objectives have been addressed: to determine the best classification method for S2, and to quantify its improvement with respect to the SPOT operational mission. To do so, four classifiers (LDA, RF, Decision Trees, K-NN) have been selected and applied to two different agricultural areas located in Valencia (Spain) and Buenos Aires (Argentina). All classifiers were tested using, on the one hand, all the S2 bands and, on the other hand, only selecting those bands from S2 closer to the four bands from SPOT. In all the cases, between 10%-50% of samples were used to train the classifier while remaining the rest for validation. As a result, a land use map was generated from the best classifier, according to the Kappa index, providing scientifically relevant information such as the area of each land use class.Keywords: classification, remote sensing, land use, Kappa Index, Sentinel-2.To cite this article: Borràs, J., Delegido, J., Pezzola, A., Pereira, M., Morassi, G., Camps-Valls, G. 2017. Land use classification from Sentinel-2 imagery. Revista de Teledetección, 48,[55][56][57][58][59][60][61][62][63][64][65][66]. https://doi.org/10.4995/raet.2017.7133 IntroducciónEl uso del suelo es un criterio socioeconómico que hace referencia a la actividad dominante que se da en un determinado lugar, e incluye usos como residencial, comercial, industrial, agrícola, etc., pudiendo estar estas categorías divididas en niveles más específicos. La cobertura del suelo es

Remote sensing evapotranspiration estimation over agricultural areas is increasingly used for irrigation management during the crop growing cycle. Different methodologies based on remote sensing have emerged for the leaf area index (LAI) and the canopy chlorophyll content (CCC) estimation, essential biophysical parameters for crop evapotranspiration monitoring. Using Sentinel-2 (S2) spectral information, this study performed a comparative analysis of empirical (vegetation indices), semi-empirical (CLAIR model with fixed and calibrated extinction coefficient) and artificial neural network S2 products derived from the Sentinel Application Platform Software (SNAP) biophysical processor (ANN S2 products) approaches for the estimation of LAI and CCC. Four independent in situ collected datasets of LAI and CCC, obtained with standard instruments (LAI-2000, SPAD) and a smartphone application (PocketLAI), were used. The ANN S2 products present good statistics for LAI (R2 > 0.70, root mean square error (RMSE) < 0.86) and CCC (R2 > 0.75, RMSE < 0.68 g/m2) retrievals. The normalized Sentinel-2 LAI index (SeLI) is the index that presents good statistics in each dataset (R2 > 0.71, RMSE < 0.78) and for the CCC, the ratio red-edge chlorophyll index (CIred-edge) (R2 > 0.67, RMSE < 0.62 g/m2). Both indices use bands located in the red-edge zone, highlighting the importance of this region. The LAI CLAIR model with a fixed extinction coefficient value produces a R2 > 0.63 and a RMSE < 1.47 and calibrating this coefficient for each study area only improves the statistics in two areas (RMSE ≈ 0.70). Finally, this study analyzed the influence of the LAI parameter estimated with the different methodologies in the calculation of crop potential evapotranspiration (ETc) with the adapted Penman–Monteith (FAO-56 PM), using a multi-temporal dataset. The results were compared with ETc estimated as the product of the reference evapotranspiration (ETo) and on the crop coefficient (Kc) derived from FAO table values. In the absence of independent reference ET data, the estimated ETc with the LAI in situ values were considered as the proxy of the ground-truth. ETc estimated with the ANN S2 LAI product is the closest to the ETc values calculated with the LAI in situ (R2 > 0.90, RMSE < 0.41 mm/d). Our findings indicate the good validation of ANN S2 LAI and CCC products and their further suitability for the implementation in evapotranspiration retrieval of agricultural areas.

The objective of this work is to evaluate the capacity of the C-band Synthetic Aperture Radar (SAR) time series imagery, acquired by the European satellite Sentinel-1 (S1), for the agriculture crop classification and its reliability to differentiate onion from sunflower, among others. The work then focused on classifying land cover in intensively cultivated agricultural regions. The study was developed in the Bonaerense Valley of the Colorado River (BVCR), Buenos Aires Province in Argentina, backed up by the field truth of 1634 field samples. In addition to the onion and sunflower crops, there are other crops present in the study area such as cereals, alfalfa, potatoes and maize, which are considered as the image background in the classification process. The field samples database was used for training and supporting a supervised classification with two machine learning algorithms—Random Forest (RF) and Support Vector Machine (SVM)—obtaining high levels of accuracy in each case. Different S1 SAR time-series features were used to assess the performance of S1 crop classification in terms of polarization VH+VV, Grey Level Co-occurrence Matrix (GLCM) image texture and a combination of both. The analysis of SAR data and their features was carried out at OBIA lot level (Object Based Image Analysis) showing an optimal strategy to counteract the effect of the residual and inherent speckle noise of the radar signal. In the process of differentiating onion and sunflower crops, the analysis of the VH+VV stack with the SVM algorithm delivered the best statistical classification results in terms of Overall Accuracy (OA) and Kappa Index, (Kp) when other crops (image background) were not considered (OA = 95.35%, Kp = 0.89). Certainly, the GLCM texture analysis derived from the S1 SAR images is a valuable source of information for obtaining very good classification results. When differentiating sunflower from onion considering also other crops present in the BVCR, the GLCM stack proved to be the most suitable dataset analyzed in this work (OA = 89.98%, Kp = 0.66 for SVM algorithm). This working methodology is applicable to other irrigated valleys in Argentina dedicated to intensive crops. There are also variables inherent to each lot, soil, crop and agricultural producer that differ according to the study area and that should be considered for each case in the future.

<p>Assessment of rural fire severity is fundamental to evaluate fire damages and to analyze recovery processes in a low-cost and efficient way. Burnt areas covering shrubs and grasslands were estimated in more than 30,000 km²  in Argentina from December 2016 to January 2017. The study area presented in this work is located in the South of the Buenos Aires province, and it covers a semiarid area with the presence of xerophilous shrubs and grasslands. This is one of the most abundant ecosystem in Central and Southern Argentina. Field campaigns were carried out over the area affected by the fire in order to georreference the burnt plots and characterized the fire severity in 5 levels. The objective of this work is to analyze the feasibility of new satellites Sentinel-2 for fire studies, as well as provide a comparison to Landsat-8 derived results, because this mission has been one of the most used in it. Pre-fire and postfire Sentinel-2 and Landsat-8 imagery were used to analyze different band combinations to compute a Normalized Difference Spectral Index (NDSI), as well as the difference of this index before and after the fire (dNDSI). Results show a significant correlation (R² =0.72 and estimation error of 0.77) between dNDSI derived from Sentinel-2 and the severity levels obtained in the field campaign using bands 8a and 12 (NIR and SWIR), the same bands as used in the Normalized Burn Ratio (NBR). Moreover, results derived from Sentinel-2 are better than results derived from Landsat-8 (R² =0.63 and estimation error of 0.92). Furthermore, it is observed that the correlation is improved when Sentinel-2 bands 6 and 5 (located in the Red-Edge region) are considered (R² =0.74 and estimation error of 0.76). An inverse correlation has been observed between the recovery of vegetation four months after the fire and the fire severity level.</p>

Earth observation offers an unprecedented opportunity to monitor intensively cultivated areas providing key support to assess fertilizer needs and crop water uptake. Routinely, vegetation traits mapping can help farmers to monitor plant development along the crop’s phenological cycle, which is particularly relevant for irrigated agricultural areas. The high spatial and temporal resolution of the Sentinel-2 (S2) multispectral instrument leverages the possibility to estimate leaf area index (LAI), canopy chlorophyll content (CCC), and vegetation water content (VWC) from space. Therefore, our study presents a hybrid retrieval workflow combining a physically-based strategy with a machine learning regression algorithm, i.e., Gaussian processes regression, and an active learning technique to estimate LAI, CCC and VWC of irrigated winter wheat. The established hybrid models of the three traits were validated against in-situ data of a wheat campaign in the Bonaerense valley, South of the Buenos Aires Province, Argentina, in the year 2020. We obtained good to highly accurate validation results with LAI: R2 = 0.92, RMSE = 0.43 m2 m−2, CCC: R2 = 0.80, RMSE = 0.27 g m−2 and VWC: R2 = 0.75, RMSE = 416 g m−2. The retrieval models were also applied to a series of S2 images, producing time series along the seasonal cycle, which reflected the effects of fertilizer and irrigation on crop growth. The associated uncertainties along with the obtained maps underlined the robustness of the hybrid retrieval workflow. We conclude that processing S2 imagery with optimised hybrid models allows accurate space-based crop traits mapping over large irrigated areas and thus can support agricultural management decisions.

No abstract

Synthetic aperture radar (SAR) data provides an appealing opportunity for all-weather day or night Earth surface monitoring. The European constellation Sentinel-1 (S1) consisting of S1-A and S1-B satellites offers a suitable revisit time and spatial resolution for the observation of croplands from space. The C-band radar backscatter is sensitive to vegetation structure changes and phenology as well as soil moisture and roughness. It also varies depending on the local incidence angle (LIA) of the SAR acquisition’s geometry. The LIA backscatter dependency could therefore be exploited to improve the retrieval of the crop biophysical variables. The availability of S1 radar time-series data at distinct observation angles holds the feasibility to retrieve leaf area index (LAI) evolution considering spatiotemporal coverage of intensively cultivated areas. Accordingly, this research presents a workflow merging multi-date S1 smoothed data acquired at distinct LIA with a Gaussian processes regression (GPR) and a cross-validation (CV) strategy to estimate cropland LAI of irrigated winter wheat. The GPR-S1-LAI model was tested against in situ data of the 2020 winter wheat campaign in the irrigated valley of Colorador river, South of Buenos Aires Province, Argentina. We achieved adequate validation results for LAI with RCV2 = 0.67 and RMSECV = 0.88 m2 m−2. The trained model was further applied to a series of S1 stacked images, generating temporal LAI maps that well reflect the crop growth cycle. The robustness of the retrieval workflow is supported by the associated uncertainties along with the obtained maps. We found that processing S1 smoothed imagery with distinct acquisition geometries permits accurate radar-based LAI modeling throughout large irrigated areas and in consequence can support agricultural management practices in cloud-prone agri-environments.

Optical Earth Observation is often limited by weather conditions such as cloudiness. Radar sensors have the potential to overcome these limitations, however, due to the complex radar-surface interaction, the retrieving of crop biophysical variables using this technology remains an open challenge. Aiming to simultaneously benefit from the optical domain background and the all-weather imagery provided by radar systems, we propose a data fusion approach focused on the cross-correlation between radar and optical data streams. To do so, we analyzed several multiple-output Gaussian processes (MOGP) models and their ability to fuse efficiently Sentinel-1 (S1) Radar Vegetation Index (RVI) and Sentinel-2 (S2) vegetation water content (VWC) time series over a dry agri-environment in southern Argentina. MOGP models not only exploit the auto-correlations of S1 and S2 data streams independently but also the inter-channel cross-correlations. The S1 RVI and S2 VWC time series at the selected study sites being the inputs of the MOGP models proved to be closely correlated. Regarding the set of assessed models, the Convolutional Gaussian model (CONV) delivered noteworthy accurate data fusion results over winter wheat croplands belonging to the 2020 and 2021 campaigns (NRMSEwheat2020 = 16.1%; NRMSEwheat2021 = 10.1%). Posteriorly, we removed S2 observations from the S1 & S2 dataset corresponding to the complete phenological cycles of winter wheat from September to the end of December to simulate the presence of clouds in the scenes and applied the CONV model at the pixel level to reconstruct spatiotemporally-latent VWC maps. After applying the fusion strategy, the phenology of winter wheat was successfully recovered in the absence of optical data. Strong correlations were obtained between S2 VWC and S1 & S2 MOGP VWC reconstructed maps for the assessment dates (R2¯wheat−2020 = 0.95, R2¯wheat−2021 = 0.96). Altogether, the fusion of S1 SAR and S2 optical EO data streams with MOGP offers a powerful innovative approach for cropland trait monitoring over cloudy high-latitude regions.