Multi-temporal SAR Research Articles

Optimal multi-temporal Sentinel-1A SAR imagery for paddy rice field discrimination; a recommendation for operational mapping initiatives

Phenology-based rice mapping has mostly employed all satellite data over the growing season. This approach is not only time consuming but also requires huge data storage and is thus unsuitable for operational initiatives. Even with the use of feature selection algorithms that minimise data redundancy, the huge amount of time in data processing prior to optimal image subset selection is unavoidable. Thus, this study seeks to identify the rice development phase(s) whose Sentinel-1A data in vertical transmitted and horizontal received (VH) and vertical transmitted and vertical received (VV) polarizations offer the most accurate rice field discrimination. Using an area in southeast China during the 2017 growing season, VH + VV data of the vegetative and maturity phases recorded the highest overall accuracy (89.3%) based on Random Forest (RF). For the agro-ecology investigated in this study, therefore, potential users of Sentinel-1A in rice mapping can ex-ante select images acquired over the crop's vegetative and maturity phases to avoid the enormous computational time and storage attendant with entire-season data. These findings would complement algorithms like Support Vector Machine (SVM) that lack feature section tools. However, replications of this study at multiple years and ecologies, and comparison with classifications obtained from optimal data selection tools (variable importance) are strongly recommended to arrive at universally applicable conclusions.

Sentinel SAR-optical fusion for crop type mapping using deep learning and Google Earth Engine

Accurate crop type mapping provides numerous benefits for a deeper understanding of food systems and yield prediction. Ever-increasing big data, easy access to high-resolution imagery, and cloud-based analytics platforms like Google Earth Engine have drastically improved the ability for scientists to advance data-driven agriculture with improved algorithms for crop type mapping using remote sensing, computer vision, and machine learning. Crop type mapping techniques mainly relied on standalone SAR and optical imagery, few studies investigated the potential of SAR-optical data fusion, coupled with virtual constellation, and 3-dimensional (3D) deep learning networks. To this extent, we use a deep learning approach that utilizes the denoised backscatter and texture information from multi-temporal Sentinel-1 SAR data and the spectral information from multi-temporal optical Sentinel-2 data for mapping ten different crop types, as well as water, soil and urban area. Multi-temporal Sentinel-1 data was fused with multi-temporal optical Sentinel-2 data in an effort to improve classification accuracies for crop types. We compared the results of the 3D U-Net to the state-of-the-art deep learning networks, including SegNet and 2D U-Net, as well as commonly used machine learning method such as Random Forest. The results showed (1) fusing multi-temporal SAR and optical data yields higher training overall accuracies (OA) (3D U-Net 0.992, 2D U-Net 0.943, SegNet 0.871) and testing OA (3D U-Net 0.941, 2D U-Net 0.847, SegNet 0.643) for crop type mapping compared to standalone multi-temporal SAR or optical data (2) optical data fused with denoised SAR data via a denoising convolution neural network (OA 0.912) performed better for crop type mapping compared to optical data fused with boxcar (OA 0.880), Lee (OA 0.881), and median (OA 0.887) filtered SAR data and (3) 3D convolutional neural networks perform better than 2D convolutional neural networks for crop type mapping (SAR OA 0.912, optical OA 0.937, fused OA 0.992).

Classification and Observed Seasonal Phenology of Broadleaf Deciduous Forests in a Tropical Region by Using Multitemporal Sentinel-1A and Landsat 8 Data

Broadleaf deciduous forests (BDFs) or dry dipterocarp forests play an important role in biodiversity conservation in tropical regions. Observations and classification of forest phenology provide valuable inputs for ecosystem models regarding its responses to climate change to assist forest management. Remotely sensed observations are often used to derive the parameters corresponding to seasonal vegetation dynamics. Data acquired from the Sentinel-1A satellite holds a great potential to improve forest type classification at a medium-large scale. This article presents an integrated object-based classification method by using Sentinel-1A and Landsat 8 OLI data acquired during different phenological periods (rainy and dry seasons). The deciduous forest and nondeciduous forest areas are classified by using NDVI (normalized difference vegetation index) from Landsat 8 cloud-free composite images taken during dry (from February to April) and rainy (from June to October) seasons. Shorea siamensis Miq. (S. siamensis), Shorea obtusa Wall. ex Blume (S. obtusa), and Dipterocarpus tuberculatus Roxb. (D. tuberculatus) in the deciduous forest area are classified based on the correlation between phenology of BDFs in Yok Don National Park and backscatter values of time-series Sentinel-1A imagery in deciduous forest areas. One hundred and five plots were selected during the field survey in the study area, consisting of dominant deciduous species, tree height, and canopy diameter. Thirty-nine plots were used for training to decide the broadleaf deciduous forest areas of the classified BDFs by the proposed method, and the other sixty-six plots were used for validation. Our proposed approach used the changes of backscatter in multitemporal SAR images to implement BDF classification mapping with acceptable accuracy. The overall accuracy of classification is about 79%, with a kappa coefficient of 0.7. Accurate classification and mapping of the BDFs using the proposed method can help authorities implement forest management in the future.

Inundation Assessment of the 2019 Typhoon Hagibis in Japan Using Multi-Temporal Sentinel-1 Intensity Images

Typhoon Hagibis passed through Japan on October 12, 2019, bringing heavy rainfall over half of Japan. Twelve banks of seven state-managed rivers collapsed, flooding a wide area. Quick and accurate damage proximity maps are helpful for emergency responses and relief activities after such disasters. In this study, we propose a quick analysis procedure to estimate inundations due to Typhoon Hagibis using multi-temporal Sentinel-1 SAR intensity images. The study area was Ibaraki Prefecture, Japan, including two flooded state-managed rivers, Naka and Kuji. First, the completely flooded areas were detected by two traditional methods, the change detection and the thresholding methods. By comparing the results in a part of the affected area with our field survey, the change detection was adopted due to its higher recall accuracy. Then, a new index combining the average value and the standard deviation of the differences was proposed for extracting partially flooded built-up areas. Finally, inundation maps were created by merging the completely and partially flooded areas. The final inundation map was evaluated via comparison with the flooding boundary produced by the Geospatial Information Authority (GSI) and the Ministry of Land, Infrastructure, Transport, and Tourism (MLIT) of Japan. As a result, 74% of the inundated areas were able to be identified successfully using the proposed quick procedure.

Detection of Forest Windstorm Damages with Multitemporal SAR Data—A Case Study: Finland

The purpose of this study was to develop methods to localize forest windstorm damages, assess their severity and estimate the total damaged area using space-borne SAR data. The development of the methods is the first step towards an operational system for near-real-time windstorm damage monitoring, with a latency of only a few days after the storm event in the best case. Windstorm detection using SAR data is not trivial, particularly at C-band. It can be expected that a large-area and severe windstorm damage may affect backscatter similar to clear cutting operation, that is, decrease the backscatter intensity, while a small area damage may increase the backscatter of the neighboring area, due to various scattering mechanisms. The remaining debris and temporal variation in the weather conditions and possible freeze–thaw transitions also affect observed backscatter changes. Three candidate windstorm detection methods were suggested, based on the improved k-nn method, multinomial logistic regression and support vector machine classification. The approaches use multitemporal ESA Sentinel-1 C-band SAR data and were evaluated in Southern Finland using wind damage data from the summer 2017, together with 27 Sentinel-1 scenes acquired in 2017 and other geo-referenced data. The stands correctly predicted severity category corresponded to 79% of the number of the stands in the validation data, and already 75% when only one Sentinel-1 scene after the damage was used. Thus, the damaged forests can potentially be localized with proposed tools within less than one week after the storm damage. In this study, the achieved latency was only two days. Our preliminary results also indicate that the damages can be localized even without separate training data.

Determining Suitable Spaceborne SAR Observations and Ground Control Points for Surface Deformation Study in Rugged Terrain With InSAR Technique

We aim to address two primary issues using the multitemporal (MT) spaceborne SAR observations and InSAR technique to monitor slope stability in rugged terrains. One is how to maximize the observation and application effectiveness based on the geometric distortion in an area of interest for determining appropriate ascending- and descending-orbit SAR observations. The other is how to optimize the selection of ground control points (GCPs) to correct orbital phase errors and reference in deriving valid ground surface displacements in the SBAS InSAR analysis. Four criteria were proposed to identify valid GCPs efficiently and effectively. The 2017 Xinmo landslide event, China, was showcased. The tear-away zone was 100% and ∼50% distorted when imaged by ascending- and descending-orbit Sentinel-1A/B SAR sensors. The descending SAR datasets were preferred. Then, we identified four spatially dispersed GCP clusters with the selection criteria using the SBAS InSAR technique and 45 descending-orbit MT SAR datasets. One GCP per cluster was chosen. With the four GCPs and SBAS analysis, we successfully detected the stable status in the landslide-free area and the prelandslide acceleration movement within the tear-away zone. Valid InSAR-derived results have been obtained. In turn, the GCP selection criteria have been validated, and the two issues resolved.

Multisensor Land Cover Classification With Sparsely Annotated Data Based on Convolutional Neural Networks and Self-Distillation

Extensive research studies have been conductedin recent years to exploit the complementarity among multi-sensor (or multi-modal) remote sensing data for prominentapplications such as land cover mapping. In order to make a step further with respect to previous studies which investigate multi-temporal SAR and optical data or multi-temporal/multi-scale optical combinations, here we propose a deep learning framework that simultaneously combine all these input sources, specifically multi-temporal SAR/optical data and fine scale optical information. Our proposal relies on a patch-based multi-branch convolutional neural network (CNN) that exploits different per source CNN encoders to deal with the specificity of the input signals. Additionally, our framework is equipped with a self-distillation strategy to boost the per source analyses. This new strategy leverages the final prediction of the multi-source framework to guide the learning of the per source CNN encoders supporting the network to learn from itself.Experiments are carried out on two real world benchmarks, namely the Reunion island (a french overseas department) and the Dordogne study site (a southwest department in France)where the annotated reference data was collected under operational constraints (sparsely annotated ground truth data). Obtained results, providing an overall classification accuracyof about 94% (resp. 88%) on theReunion island(resp. theDordogne) study site. These findings highlight the effectivenessof our framework based on Convolutional Neural Networks andself-distillation to combine heterogeneous multi-sensor remotesensing data and confirm the benefit of multi-modal analysisfor downstream tasks such as land cover mapping.

Multitemporal SAR interferometry for monitoring of ground deformations caused by hydrocarbon production in an arid environment: Case studies from the Sultanate of Oman

Time-lapse interferometric synthetic aperture radar (InSAR) remote sensing methods of surface deformation have proven their use in desert environments. The data are acquired frequently without the need to send personnel or equipment into the field. The quality and accuracy of the data is very high. The spatial resolution of the data is excellent and matches that of surface seismic data. These characteristics make the data well suited for a variety of time-lapse monitoring tasks. In this study, we describe the accuracy of the InSAR technique relative to other measurements such as the global positioning system and precise leveling (acquired at known stable locations). We illustrate two case studies of differing natures. In one case, gas production leads to reservoir compaction, which is tracked as surface subsidence with time using frequent InSAR data. The results are used to map zones of increased deformation and identify areas with localized changes. These insights are being used to influence decisions on new wells and well interventions, to provide support for management of facility integrity, and to advise building code and material selection that can withstand the expected rate of deformation. The second case of a shallow steam flood illustrates the use of InSAR data to identify areas of surface uplift following thermal expansion of the reservoir. These data are also used to support the monitoring of the steam chamber growth and confinement in the reservoir. The information from InSAR will become more valuable for reservoir management when the steam chamber matures and conventional downhole data acquisition consequently becomes challenging. In summary, oil and gas fields located in arid environments lend themselves well to remote sensing using the InSAR technique because (1) they are sizeable (from tens to hundreds of square kilometers); (2) they are free from vegetation, snow cover, and most atmospheric distortions, although cloud and pollution can affect the data quality; and (3) they benefit from highly repeatable long-term regular monitoring.

Comment on “Pre-Collapse Space Geodetic Observations of Critical Infrastructure: The Morandi Bridge, Genoa, Italy” by Milillo et al. (2019)

We present in this comment a Multi-Temporal SAR Interferometry (MT-InSAR) analysis showing that the results published by Milillo et al. (2019) in the Remote Sensing Journal, presenting the evidence of space geodetic observations relevant to displacements occurring before the collapse of the Morandi Bridge, happened in Genova (Italy) on the 14 August 2018, are questionable. In particular, we focus on the InSAR results obtained by Milillo et al. (2019) by processing the 3 m × 3 m resolution COSMO-SkyMed (CSK) data collected from ascending and descending orbits on the area of interest. These results, thanks to the high spatial resolution and the short revisit time characterizing this multi-orbit SAR dataset, represent the cornerstone of their analysis. The main findings of their study allow Milillo et al. to conclude that the InSAR processing of this COSMO-SkyMed dataset reveals the increased deformation magnitude over time of points located near the strands of the deck next to the collapsed pier, between 12 March 2017 and August 2018. In this comment, we show the results obtained by the IREA-CNR SAR team after processing the same ascending and descending CSK dataset, but by using two alternative and independent processing techniques: the Small BAseline Subset (SBAS) and the Advanced Tomographic SAR (TomoSAR) approaches, respectively. Our analysis shows that, although both the SBAS and the TomoSAR analyses allow achieving denser coherent pixel maps relevant to the Morandi bridge, nothing of the pre-collapse large displacements reported in Milillo et al. (2019) appears in our results, leading us to deeply disagree with the findings of their InSAR analysis.

Monthly Deforestation Monitoring with Sentinel-1 Multi-temporal Signatures and InSAR Coherences

<p class="p1">Sentinel-1 interferometric time-series allow for the accurate retrieval of the target’s temporal decorrelation and, therefore, the inversion of land cover information and its temporal monitoring. This paper describes the development of an observation scenario for monitoring monthly deforestation over the Amazon rainforest, which relies on the use of radar for overcoming the physical limitations of optical sensors caused by the presence of cloud coverage. Specifically, we implement a classification scheme that exploits multi-temporal SAR features, such as backscatter, spatial textures, and interferometric parameters, to map forested areas. Distinct forest maps are generated for consecutive months and further processed to detect deforestation phenomena and map clear-cuts evolution. The obtained results are validated by selecting cloud-free Sentinel-2 multispectral data on the selected area and acquired during the same observation time.</p>

MAPPING EXTENSION AND MAGNITUDE OF CHANGES INDUCED BY CYCLONE IDAI WITH MULTI-TEMPORAL LANDSAT AND SAR IMAGES

Abstract. In this paper it is described a study case of a rapid assessment of change detections for post-cyclone Idai vegetated damage and flood extension estimation by fusion of multi-temporal Landsat and sentinel-1 SAR images. For automated change detection, after disasters, many algorithms have been proposed. To visualize the changes induced by cyclone we tested and compared two automated change detection techniques namely: Principal Components Analysis (PCA), Normalized Difference Vegetation Index (NDVI) and image segmentation. With the image segmentation of multispectral and SAR images, it was possible to visualize the extension of the wet area. For this specific application, PCA was identified as the optimal change detection indicator than NDVI. This study suggested that image segmentation, principal components analysis, and normalized difference vegetation index can be used for change detection of surface water due to flood and disasters especially in prone countries like Mozambique.

Adaptive Multilooking of Multitemporal Differential SAR Interferometric Data Stack Using Directional Statistics

In this article, we present an innovative space–time adaptive multilooking technique that operates on a sequence of multitemporal, differential synthetic aperture radar interferograms. The developed approach relies on the application of the fundamentals of directional statistics theory. At variance with other methods that identify the set of statistically homogenous pixels (SHPs) within a multilooking (complex averaging) window based on the statistics of the single-look-complex (SLC) SAR images, the proposed method is exclusively based on the analysis of the multitemporal sequence of full resolution DInSAR interferograms. The SHPs are then used to generate spatially adaptive multilooked interferograms both at the native, full-scale grid of the SLC images and at the multilooked resolution scale. The algorithm is effective and simple to implement, only requiring the availability of a sequence of full-scale differential SAR interferometry (DInSAR) interferograms. The interferograms can then be used to generate ground displacement time-series through advanced multitemporal interferometric SAR (MTInSAR) approaches. Experimental results obtained by applying the adopted technique to two SAR data sets acquired at X- and L-band, respectively, demonstrate the validity of the developed method.

Deep Learning Applications on Multitemporal SAR (Sentinel-1) Image Classification Using Confined Labeled Data: The Case of Detecting Rice Paddy in South Korea

The applicability of deep learning to remote sensing is rapidly increasing in accordance with the improvement in spatiotemporal resolution of satellite images. However, unlike satellite images acquired in near-real-time over wide areas, there are limited amount of labeled data used for model training. In this article, three kinds of deep learning applications—data augmentation, semisupervised classification, and domain-adapted architecture—were tested in an effort to overcome the limitation of insufficient labeled data. Among the diverse tasks that can be used for classification, rice paddy detection in South Korea was performed for its ability to fully utilize the advantages of deep learning and high spatiotemporal image resolution. In the process of designing each application, the domain knowledge of remote sensing and rice phenology was integrated. Then, all possible combinations of the three applications were examined and evaluated with pixel-based comparisons in various environments and city-level comparisons using national statistics. The results of this article indicated that all combinations of the applications can contribute to increase classification performance, even though the uncertainty involved in imitating or utilizing unlabeled data remains. As the effectiveness of the proposed applications was experimentally confirmed, enhancement in the applicability of deep learning was expected in various remote sensing areas. In particular, the proposed applications would be significant when they are applied to a wide range of study areas and high-resolution images, as they tend to require a large amount of learning data from diverse environments, owing to high intra-class heterogeneity.

Surface Roughness Estimation in the Orog Nuur Basin (Southern Mongolia) Using Sentinel-1 SAR Time Series and Ground-Based Photogrammetry

This study demonstrates an application-oriented approach to estimate area-wide surface roughness from Sentinel-1 time series in the semi-arid environment of the Orog Nuur Basin (southern Mongolia) to support recent geomorphological mapping efforts. The relation of selected mono- and multi-temporal SAR features and roughness is investigated by using an empirical multi-model approach and selected 1D and 2D surface roughness indices. These indices were obtained from 48 high-resolution ground-based photogrammetric digital elevation models, which were acquired during a single field campaign. The analysis is backed by a time series analysis, comparing Sentinel-1 features to temporal-corresponding observations and reanalysis datasets on soil moisture conditions, land surface temperature, occurrence of precipitation events, and presence and development of vegetation. Results show that Sentinel-1 features are hardly sensitive to the changing surface conditions over none to sparsely vegetated land, indicating very dry conditions throughout the year. Consequently, surface roughness is the dominating factor altering SAR intensity. The best correlation is found for the combined surface roughness index Z-Value (ratio between the root mean square height and the correlation length) and the mean summer VH intensity with an r2 coefficient of 0.83 and an Root-Mean-Square Error of 0.032.

Evaluation of Water-Saving Methodologies for Rice Farming Using a Distributed Long-Term Rainfall-Runoff Model Incorporated Agricultural Ponds Detected by Multi-Temporal SAR Images in the Ibague Watershed, Colombia

Evaluation of Water-Saving Methodologies for Rice Farming Using a Distributed Long-Term Rainfall-Runoff Model Incorporated Agricultural Ponds Detected by Multi-Temporal SAR Images in the Ibague Watershed, Colombia

National Scale Surface Deformation Time Series Generation through Advanced DInSAR Processing of Sentinel-1 Data within a Cloud Computing Environment

We present an automatic pipeline implemented within the Amazon Web Services (AWS) Cloud Computing platform for the interferometric processing of large Sentinel-1 (S1) multi-temporal SAR datasets, aimed at analyzing Earth surface deformation phenomena at wide spatial scale. The developed processing chain is based on the advanced DInSAR approach referred to as Small BAseline Subset (SBAS) technique, which allows producing, with centimeter to millimeter accuracy, surface deformation time series and the corresponding mean velocity maps from a temporal sequence of SAR images. The implemented solution addresses the aspects relevant to i) S1 input data archiving; ii) interferometric processing of S1 data sequences, performed in parallel on the AWS computing nodes through both multi-node and multi-core programming techniques; iii) storage of the generated interferometric products. The experimental results are focused on a national scale DInSAR analysis performed over the whole Italian territory by processing 18 S1 slices acquired from descending orbits between March 2015 and April 2017, corresponding to 2612 S1 acquisitions. Our analysis clearly shows that an effective integration of advanced remote sensing methods and new ICT technologies can successfully contribute to deeply investigate the Earth System processes and to address new challenges within the Big Data EO scenario.

Facies detection of surge-type glaciers, Karakoram Himalaya using Sentinel-1 images

Glacier facies or radar glacier zones (RGZs) of the surge-type glaciers of the Karakoram Himalaya were classified based on the differences in the backscatter values utilizing multi-temporal SAR data from Sentinel-1 satellite of the years 2018 and 2019. Distinct zones of glaciers such as dry snow zones, percolation zones, wet snow zone, bare ice zone and debris-covered area can be discernible on the false-color composite image made from the early summer-late summer-winter (R-G-B) of SAR images obtained from Sentinel-1. The result suggests an interesting blended pattern of facies distribution on surge-type glaciers showing the impression of accumulation zones at the lower most parts of the glaciers contrary to typical non surge-type glaciers. We found that the facies of the surge-type glaciers of the Karakoram Himalaya did not follow the relationship with the elevation in terms of their distribution on the surface of glaciers. We also compared our results with the surface temperature of the glaciers derived from the thermal band of Landsat ETM+.

Monitoring Agricultural Fields Using Sentinel-1 and Temperature Data in Peru: Case Study of Asparagus (Asparagus officinalis L.)

This paper presents the analysis and a methodology for monitoring asparagus crops from remote sensing observations in a tropical region, where the local climatological conditions allow farmers to grow two production cycles per year. We used the freely available dual-polarisation GRD data provided by the Sentinel-1 satellite, temperature from a ground station and ground truth from January to August of 2019 to perform the analysis. We showed how particularly the VH polarisation can be used for monitoring the canopy formation, density and the growth rate, revealing connections with temperature. We also present a multi-output machine learning regression algorithm trained on a rich spatio-temporal dataset in which each output estimates the number of asparagus stems that are present in each of the pre-defined crop phenological stages. We tested several scenarios that evaluated the importance of each input data source and feature, with results that showed that the methodology was able to retrieve the number of asparagus stems in each crop stage when using information about starting date and temperature as predictors with coefficients of determination ( R 2 ) between 0.84 and 0.86 and root mean squared error (RMSE) between 2.9 and 2.7. For the multitemporal SAR scenario, results showed a maximum R 2 of 0.87 when using up to 5 images as input and an RMSE that maintains approximately the same values as the number of images increased. This suggests that for the conditions evaluated in this paper, the use of multitemporal SAR data only improved mildly the retrieval when the season start date and accumulated temperature are used to complement the backscatter.

Deep learning networks for selection of measurement pixels in multi-temporal SAR interferometric processing

In multi-temporal SAR interferometry (MT-InSAR), persistent scatterer (PS) pixels are used to estimate geophysical parameters, essentially deformation. Conventionally, PS pixels are selected based on the estimated noise present in the spatially uncorrelated phase component along with look-angle error in a temporal interferometric stack. In this study, two deep learning architectures, namely convolutional neural network for interferometric semantic segmentation (CNN-ISS) and convolutional long short term memory network for interferometric semantic segmentation (CLSTM-ISS), based on learning spatial and spatio-temporal behaviour, respectively, were proposed for selection of PS pixels. These networks were trained to relate the interferometric phase history to its classification into phase stable (PS pixels) and phase unstable (non-PS pixels) measurement pixels using ~10,000 real world interferometric patch images of different study sites containing man-made objects, forests, vegetation, uncropped land, water bodies, and areas affected by lengthening, foreshortening, layover and shadowing. The networks were trained using training labels obtained from the Stanford method for Persistent Scatterer Interferometry (StaMPS) algorithm. However, pixel selection results, evaluated using a combination of R-index, Similar Time Series Interferometric Pixel (STIP) maps and a classified image of the test dataset, reveal that CLSTM-ISS estimates improved the classification of PS and non-PS pixels as compared to those of StaMPS and CNN-ISS. The predicted results show that CLSTM-ISS reached an accuracy of 93.50%, higher than that of CNN-ISS (89.21%). CLSTM-ISS also improved the density of reliable PS pixels compared to StaMPS and CNN-ISS. Further, the architecture outperformed StaMPS, and is expected to compete with other MT-InSAR algorithms in terms of computational efficiency.

Mapping wetland using the object-based stacked generalization method based on multi-temporal optical and SAR data

Wetland ecosystems have experienced dramatic challenges in the past few decades due to natural and human factors. Wetland maps are essential for the conservation and management of terrestrial ecosystems. This study is to obtain an accurate wetland map using an object-based stacked generalization (Stacking) method on the basis of multi-temporal Sentinel-1 and Sentinel-2 data. Firstly, the Robust Adaptive Spatial Temporal Fusion Model (RASTFM) is used to get time series Sentinel-2 NDVI, from which the vegetation phenology variables are derived by the threshold method. Subsequently, both vertical transmit-vertical receive (VV) and vertical transmit-horizontal receive (VH) polarization backscatters (σ0 VV, σ0 VH) are obtained using the time series Sentinel-1 images. Speckle noise inherent in SAR data, resulting in over-segmentation or under-segmentation, can affect image segmentation and degrade the accuracies of wetland classification. Therefore, we segment Sentinel-2 multispectral images to delineate meaningful objects in this study. Then, in order to reduce data redundancy and computation time, we analyze the optimal feature combination using the Sentinel-2 multispectral images, Sentinel-2 NDVI time series, phenological variables and other vegetation index derived from Sentinel-2 multispectral images, as well as time series Sentinel-1 backscatters at the object level. Finally, the stacked generalization algorithm is utilized to extract the wetland information based on the optimal feature combination in the Dongting Lake wetland. The overall accuracy and Kappa coefficient of the object-based stacked generalization method are 92.46% and 0.92, which are 3.88% and 0.04 higher than that using the pixel-based method. Moreover, the object-based stacked generalization algorithm is superior to single classifiers in classifying vegetation of high heterogeneity areas.