Search

Incorporating optical and polarimetric synthetic-aperture radar (SAR) data to estimate impervious surface is useful but challenging due to their different geometric imaging mechanism and the high diversity of urban land covers. The recent development of deep convolutional networks (DCN) opens a promising opportunity by automatically extracting the deep features from both data sets. In this study, a small-patched DCN (SDCN) was designed to estimate the impervious surface from optical and SAR data. Benchmark methods, e.g., GoogLeNet, VGG16, ResNet50, and the support vector machine were employed for comparison. Two study sites in the most complex metropolitan of China, the Guangdong-Hong Kong-Macau Greater Bay Area, were selected to assess the proposed method. Experimental results indicated the effectiveness of proposed SDCN with a better accuracy outperforming other benchmark methods. Furthermore, we found that 60%–80% of training samples performed comparably with the whole training set, indicating that a large number of training samples may not be necessary in all cases, depending on the settings of some factors (e.g., number of epochs). Generally, SDCN appears more suitable than other methods in terms of combining the optical and SAR data and improved the accuracy of estimating impervious surface. [ABSTRACT FROM AUTHOR]

Numerous studies on environmental modeling and ecological process lay emphasis on the fundamental information of the impervious surface area (ISA). However, accurate ISA extraction from high-resolution satellite images remains challenging due to both the high diversity of land covers and shadow effects from tall buildings and trees. To address the problem, a discriminative Optical-SAR-LiDAR dictionary sparse representation classification (OSLD-SRC) method was proposed using high-resolution WorldView-2, GeoEye-1, TerraSAR-X, and airborne LiDAR data. First, it used multisource data and fuzzy samples by low-pass filtering (LPF) to solve the problem of roads and buildings misclassification; second, it learned the Optical-SAR-LiDAR dictionary for nonshadow and shadow classes, and then used discriminative sparse coding method for classification to reduce the shadow effects and improve the ISA extraction accuracy. Experimental results demonstrated the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Urban impervious surfaces have been recognized as a significant indicator for various environmental and socio-economic studies. There is an increasingly urgent demand for timely and accurate monitoring of the impervious surfaces with satellite technology from local to global scales. In the past decades, optical remote sensing has been widely employed for this task with various techniques. However, there are still a range of challenges, e.g. handling cloud contamination on optical data. Therefore, the Synthetic Aperture Radar (SAR) was introduced for the challenging task because it is uniquely all-time- and all-weather-capable. Nevertheless, with an increasing number of SAR data applied, the methodology used for impervious surfaces classification remains unchanged from the methods used for optical datasets. This shortcoming has prevented the community from fully exploring the potential of using SAR data for impervious surfaces classification. We proposed a new scheme that is comparable to the well-known and fundamental Vegetation-Impervious surface-Soil (V-I-S) model for mapping urban impervious surfaces. Three scenes of fully polarimetric Radsarsat-2 data for the cities of Shenzhen, Hong Kong and Macau were employed to test and validate the proposed methodology. Experimental results indicated that the overall accuracy and Kappa coefficient were 96.00% and 0.8808 in Shenzhen, 93.87% and 0.8307 in Hong Kong and 97.48% and 0.9354 in Macau, indicating the applicability and great potential of the new scheme for impervious surfaces classification using polarimetric SAR data. Comparison with the traditional scheme indicated that this new scheme was able to improve the overall accuracy by up to 4.6% and Kappa coefficient by up to 0.18. [ABSTRACT FROM AUTHOR]

Urban land covers (ULC) are essential data for numerous studies of urban landscape ecology performed on various scales. Nevertheless, it remains difficult to obtain accurate and timely ULC information. This study presents a methodological framework for fusing optical and synthetic aperture radar (SAR) data at the pixel level with manifolds to improve ULC classification. Three typical manifold learning models, namely, ISOMAP, Local Linear Embedding (LLE) and principle component analysis (PCA), were employed, and their results were compared. SPOT-5 data were used as optical data to be fused with three different SAR datasets. Experimental results showed that 1) the most useful information of the optical and SAR data were included in the manifolds with intrinsic dimensionality, while various ULC classes were distributed differently throughout the feature spaces of manifolds derived from different learning methods; 2) in certain cases, ISOMAP performed comparably to PCA, but PCA generally performed the best out of all the study cases, yielding the best producer's and user's accuracy of all ULC classes and requiring the least amount of time to build the machine learning models; and 3) the LLE-derived manifolds yielded the lowest accuracy, primarily by confusing bare soils with dark impervious surfaces and vegetation. These results indicate the effectiveness of the new manifold technology to fuse optical and SAR data at the pixel level for improving ULC classification, which can be applied in practice to support the accurate analysis of urban landscape. [ABSTRACT FROM AUTHOR]

The estimation of chlorophyll-a (chl-a) concentration remains a great challenge in turbid waters due to their complex optical conditions. To improve chl-a estimation, this study aims to determine whether combined use of polarimetric synthetic-aperture radar (SAR) data has potential for improving the chl-a estimation from hyperspectral sensing reflectance for turbid waters such as those found in Lake Taihu, China. In situ measurements of hyperspectral reflectance data and water samples were collected over the lake corresponding to ENVISAT ASAR data. Semiempirical (two-band and three-band models) and empirical [multiple linear regression (MLR) and multilayer perceptron network (MLP)] models are compared to estimate the chl-a concentration from in situ hyperspectral reflectance and SAR data. The results show that there is a general underestimation of chl-a for concentrations higher than 26 ug/L, which is probably caused by the large spatial variation of chl-a in the study area. The results also demonstrate that the MLR model performs in a more stable manner than the MLP network does, while MLP underestimates low and high areas of chl-a concentrations in the lake. On the other hand, due to the availability of one scenic SAR data on the same day, our results show that the additional use of SAR data improved chl-a estimation very slightly in this case study, although the performance of vertical/vertical polarization SAR data was better than that of horizontal/horizontal polarization data in chl-a estimation. Potential future work in this subject could explore other measures of mutual information between SAR and hyperspectral optical data beyond the correlation and regression techniques described. Therefore, it is still necessary to apply more SAR data in varied turbid waters in the near future to determine how SAR data can be useful in the improvement of chl-a estimation. [ABSTRACT FROM PUBLISHER]

Integrating synthetic aperture radar (SAR) and optical data to improve urban land cover classification has been identified as a promising approach. However, which integration level is the most suitable remains unclear but important to many researchers and engineers. This study aimed to compare different integration levels for providing a scientific reference for a wide range of studies using optical and SAR data. SAR data from TerraSAR-X and ENVISAT ASAR in both WSM and IMP modes were used to be combined with optical data at pixel level, feature level and decision levels using four typical machine learning methods. The experimental results indicated that: 1) feature level that used both the original images and extracted features achieved a significant improvement of up to 10% compared to that using optical data alone; 2) different levels of fusion required different suitable methods depending on the data distribution and data resolution. For instance, support vector machine was the most stable at both the feature and decision levels, while random forest was suitable at the pixel level but not suitable at the decision level. 3) By examining the distribution of SAR features, some features (e.g., homogeneity) exhibited a close-to-normal distribution, explaining the improvement from the maximum likelihood method at the feature and decision levels. This indicated the benefits of using texture features from SAR data when being combined with optical data for land cover classification. Additionally, the research also shown that combining optical and SAR data does not guarantee improvement compared with using single data source for urban land cover classification, depending on the selection of appropriate fusion levels and fusion methods. [ABSTRACT FROM AUTHOR]

Numerous studies have demonstrated the benefits of fusing optical and synthetic aperture radar (SAR) data for urban impervious surfaces estimation. Given the increasing availability of SAR satellite data, the synergized use of optical and SAR data is increasingly important. However, previous studies have focused primarily on a single scale. This study aims to evaluate the impervious surfaces estimation using different combinations of optical and SAR data, as well as to quantify the scale effects of combining optical and SAR data under various resolutions from 30 to 1050 m. Six combinations of optical and SAR data in three cities, Shenzhen, Mumbai, and São Paulo, were compared. Experimental results illustrated the effectiveness of using texture and SAR data in improving impervious surface estimation with producer's accuracy improvement of 5%, 0%, and 4% in Shenzhen, Mumbai, and São Paulo at 30 m. The correlation coefficient assessment between the estimated and referenced impervious surfaces at various scales demonstrated that TerraSAR-X data generally outperformed ENVISAT advanced SAR of impervious surface estimation. Estimation accuracy of impervious surfaces in Shenzhen had an improvement of 4% to 5%. Moreover, the texture information of SAR data was observed to provide positive contributions to the impervious surfaces estimation in three study cases at different scales. [ABSTRACT FROM PUBLISHER]

Multisource remote sensing data have been widely used to improve land-cover classifications. The combination of synthetic aperture radar (SAR) and optical imagery can detect different land-cover types, and the use of genetic algorithms (GAs) and support vector machines (SVMs) can lead to improved classifications. Moreover, SVM kernel parameters and feature selection affect the classification accuracy. Thus, a GA was implemented for feature selection and parameter optimization. In this letter, a GA-SVM algorithm was proposed as a method of classifying multifrequency RADARSAT-2 (RS2) SAR images and Thaichote (THEOS) multispectral images. The results of the GA-SVM algorithm were compared with those of the grid search algorithm, a traditional method of parameter searching. The results showed that the GA-SVM algorithm outperformed the grid search approach and provided higher classification accuracy using fewer input features. The images obtained by fusing RS2 data and THEOS data provided high classification accuracy at over 95%. The results showed improved classification accuracy and demonstrated the advantages of using the GA-SVM algorithm, which provided the best accuracy using fewer features. [ABSTRACT FROM PUBLISHER]

With the increasing applications of multisource remote sensing, the synergistic use of optical and Synthetic Aperture Radar (SAR) remote sensing images is becoming increasingly popular. Although numerous approaches have been proposed to register satellite images between two optical images or between two SAR images, these approaches cannot be directly applied to register optical and SAR images due to the differences of imaging mechanism between optical and SAR remote sensing. Therefore, the co-registration between optical and SAR images remains challenging, especially in urban areas where surface geometric configuration is much more complex. This study aims to propose a fast and automatic approach for optical and SAR data co-registration based on an assumption of affine transformation between the two images. With a set of initial homologous points generated automatically, an optimization is applied towards the initial homologous points on the SAR images subject to an affine transformation model and a given threshold of root mean square errors (RMSEs) for the co-registration. Three various data sets of optical and SAR images, including SPOT-5 and ENVISAT ASAR, Landsat TM and ENVISAT ASAR, and SPOT-5 and ALOS PALSAR, were employed to test and validate the effectiveness of the proposed method. Experimental results demonstrate the promising effectiveness of the proposed approach with a significant reduction of RMSE compared with the initial homologous points. The proposed approach is especially applicable in urban areas where the uncertain errors of homologous points in SAR images are higher due to the speckles and foreshortening, layover and shadows from tall buildings. [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) data can provide complementary information to improve the mapping of urban impervious surfaces. However, most studies have focused on using only single polarization SAR data. This paper presents a comparative study on the combined use of multispectral optical data and dual polarization SAR data to identify urban impervious surfaces. The experimental results using SPOT-5, TerraSAR-X and ALOS PALSAR data were consistent compared with our previous results using single polarization SAR data. The two-fold result showed that polarimetric SAR images were generally superior to single polarization SAR data for extracting impervious surface areas, although not every individual polarimetric feature could provide a positive result for impervious surfaces mapping. Compared with using only optical and SAR data, the separate HH and HV polarization data improved the accuracy of the results. The incorporation of both Entropy and Alpha features also improved the accuracy. However, the HH/HV ratio and the separate use of coherence did not provide positive results. Noticeably, a combination of all of the dual-polarimetric SAR features was capable of obtaining the best accuracy, with an improvement of approximately 3.5% compared with that of only using SPOT-5 images. This result indicates the superiority of dual-polarimetric SAR data over single polarimetric SAR data for the mapping of urban impervious surfaces. [ABSTRACT FROM AUTHOR]

Land use/land cover (LULC) classification using optical and synthetic aperture radar (SAR) remote sensing images is becoming increasingly important to produce more accurate LULC products. As an important step, feature normalization techniques have been studied by the areas of pattern recognition. Nevertheless, because of the totally different imaging mechanisms of optical and SAR sensors, most of the existing normalization approaches are not suitable for optical and SAR data fusion. Moreover, whether normalization is a significant step remains unclear regarding optical and SAR fusion. Taking the Satellite Pour l'Observation de la Terre (SPOT-5) and the Advanced Land Observing Satellite (ALOS)/Phased Array type L-band SAR (PALSAR) (HH and HV polarizations) as the optical and SAR data, this letter aims to evaluate the impact of feature normalization. Experimental results indicated that feature normalization is not necessarily significant depending on fusion methods. For instance, distribution-dependent classifiers (e.g., a maximum likelihood classifier) are independent of feature normalization; thus, it has no impact on the results when using these classifiers. Moreover, advanced classifiers (e.g., a support vector machine) with built-in normalization are also not influenced by feature normalization. In contrast, a minimum distance classifier and an artificial neural network (ANN) depend on the input values of optical and SAR features and thus can be influenced by feature normalization. However, our experiments showed a fluctuation in classification accuracy using an ANN with normalized features. Therefore, more experiments are required to investigate the optimal normalization approaches for the optical and SAR images when using an ANN as the fusion method. [ABSTRACT FROM PUBLISHER]

An improved reconstruction algorithm is proposed for compact polarimetric (CP) synthetic aperture radar (SAR) on oil spill detection. Based on the differences in statistical behavior between open and oil-covered sea surfaces, the proposed algorithm can iteratively reconstruct quad-pol SAR images from CP SAR data. During the experiment, it outperformed two existing compact SAR reconstruction algorithms in terms of both statistical and information theoretical analysis. [ABSTRACT FROM PUBLISHER]

Abstract: Accurate mapping of urban impervious surfaces is important but challenging due to the diversity of urban land covers. This study presents an effort to synergistically combine optical and SAR data to improve the mapping of impervious surfaces. Three pairs of optical and SAR images, Landsat ETM+ and ENVISAT ASAR, SPOT-5 and ENVISAR ASAR, and SPOT-5 and TerraSAR-X, were selected in three study areas to validate the effectiveness of the methods in this study. The potential of Random Forest (RF) was evaluated with parameter optimization for combining the optical and SAR images. Experiment results demonstrate some interesting findings. Firstly, the built-in out-of-bag (OOB) error is insufficient for accuracy assessment, and an assessment with additional reference data is required for combining optical and SAR images using RF. Secondly, the optimal number of variables (m) for splitting the decision tree nodes in RF should be some different from the principles reported previously, and an empirical relationship was given for determining the parameter m. Thirdly, the optimal number of decision trees (T) in RF is not sensitive to the resolutions and sensor types of optical and SAR images, and the optimal T in this study is 20. Fourthly, the combined use of optical and SAR images by using RF is effective to improve the land cover classification and impervious surface estimation, by reducing the confusions between bright impervious surface and bare soil and dark impervious surface and bare soil, as well as shaded area and water surface. Even though the easily-confused land classes tend to be different in different resolutions of images, the effectiveness of combining optical and SAR images is consistent. This improvement is more significant when combing lower resolution optical and SAR images. The conclusions of this study could serve as an important reference for further applications of optical and SAR images, and as a potential reference for the applications of RF to the fusion of other multi-source remote sensing data. [Copyright &y& Elsevier]

Estimation of impervious surface area (ISA) from satellite images is crucial towards many environmental issues. However, many difficulties remain to be addressed, especially in cloudy areas. In this article, the potential of Advanced Synthetic Aperture Radar (ASAR) data is evaluated for mapping ISA. First, the spectral mixture analysis is employed to estimate ISA from Enhanced Thematic Mapper Plus (ETM+) image. Second, the ISA is extracted from ASAR data according to the characteristics of synthetic aperture radar. Third, a decision fusion is used to fuse the results from ETM+ and ASAR images. Accuracy assessment shows that by just using ASAR image, the accuracy (overall accuracy: 0.6815; kappa coefficient: 0.3557) is lower than that using only ETM+ image (overall accuracy: 0.7450; kappa coefficient: 0.4806). However, the fused result (overall accuracy: 0.7612; kappa coefficient: 0.5159) gives a better accuracy than using single data source, indicating that ASAR data are able to provide complementary information on ISA, since microwave works on all-weather conditions. [ABSTRACT FROM AUTHOR]

Abstract: The estimation of impervious surface area (ISA) is becoming increasingly important because of its environmental and socio-economic significance. However, accurate ISA estimation remains challenging due to the diversity of impervious materials, as well as the occurrence of clouds in subtropical humid areas. In order to address these challenges and provide an accurate estimation of ISA in cloudy areas, it is advantageous to use both optical and microwave remote sensing which can penetrate cloud coverage. Our study aims to conduct a comprehensive comparison between these two data sources and between different methods for mapping ISA. Both the classification results and accuracy assessment provide a better understanding about the differences between Landsat ETM+ and ENVISAT ASAR images and between artificial neural network (ANN) and support vector machine (SVM) classifier for estimating the impervious surfaces. The comparison demonstrates that ETM+ images alone provide a better ISA estimation (OA: about 90%; Kappa: about 0.88) than the estimation from ASAR images alone (OA: about 85%; Kappa: about 0.77). Additionally, the experiment indicates that SVM should be a better choice for ISA estimation using Landsat ETM+ images, while ANN turns out to be more sensitive to the confusion between dry soils and bright impervious surfaces, and between shade and dark impervious surfaces. For ENVISAR ASAR images, ANN gets a better result with higher accuracy, while the SVM classifier produces more noise and has some edge effects. [Copyright &y& Elsevier]

Urban land cover (ULC) serves as fundamental environmental information for urban studies, while accurate and timely ULC mapping remains challenging due to cloud contamination in tropical and subtropical areas. Synthetic aperture radar (SAR) has excellent all-weather working capability to overcome the challenge, while optical SAR data fusion is often required due to the limited land surface information provided by SAR. However, the mechanism by which SAR can compensate optical images, given the occurrence of clouds, in order to improve the ULC mapping, remains unexplored. To address the issue, this study proposes a framework, through various sampling strategies and three typical supervised classification methods, to quantify the ULC classification accuracy using optical and SAR data with various cloud levels. The land cover confusions were investigated in detail to understand the role of SAR in distinguishing land cover under different types of cloud coverage. Several interesting experimental results were found. First, 50% cloud coverage over the optical images decreased the overall accuracy by 10–20%, while the incorporation of SAR images was able to improve the overall accuracy by approximately 4%, by increasing the recognition of cloud-covered ULC information, particularly the water bodies. Second, if all the training samples were not contaminated by clouds, the cloud coverage had a higher impact with a reduction of 35% in the overall accuracy, whereas the incorporation of SAR data contributed to an increase of approximately 5%. Third, the thickness of clouds also brought about different impacts on the results, with an approximately 10% higher reduction from thick clouds compared with that from thin clouds, indicating that certain spectral information might still be available in the areas covered by thin clouds. These findings provide useful references for the accurate monitoring of ULC over cloud-prone areas, such as tropical and subtropical cities, where cloud contamination is often unavoidable. [ABSTRACT FROM AUTHOR]

The area, distribution, and temporal dynamics of anthropogenic impervious surface (AIS) at large scale are significant for environmental, ecological and socio-economic studies. Remote sensing has become an important tool for monitoring large scale AIS, while it remains challenging for accurate extraction of AIS using optical datasets alone due to the high diversity of land covers over large scale. Previous studies indicated the complementary use of synthetic aperture radar (SAR) to improve the AIS estimation, while most of them were limited to local and small scales. The potential of SAR for large scale AIS mapping is still uncertain and underexplored. In this study, first, a machine learning framework incorporating both optical and SAR data based on Google Earth Engine platform was developed for mapping and analyzing the annual dynamics of AIS in China. Feature-level fusion for SAR and optical data across large scale was tested applicable considering the backscattering coefficients, texture measures and spectral characteristics. Improved accuracy (averaged 2% increased overall accuracy and averaged 4% increased Kappa coefficient) and better delineation between the bright impervious surface and bare land was observed comparing with using optical data alone. Second, comprehensive assessment was conducted using high-resolution samples from Google Earth, census data from China Statistic Yearbook and benchmark datasets from the GlobeLand30 and GHSL, demonstrating the feasibility and reliability of the proposed method and results. Last but not the least, we analyzed the spatial and temporal patterns of AIS in China from national, regional and provincial levels. • We incorporated SAR and optical images for large scale impervious surfaces mapping. • Results were validated with high resolution samples, census and benchmark data. • SAR generally improved the overall accuracy by 2%. • SAR better delineated the impervious surface from bare land. [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) remote sensing technology has the advantage of all-weather observation and can acquire time-series images with crop growth period, which has great potential for applications such as crop phenology analysis. However, available studies primarily focus on conducting statistical and crop growth analyses based on the polarization or backscatter intensities of SAR images, and the exploration of polarization scattering information in SAR images is not sufficient. To comprehensively reflect the polarization characteristics and scattering mechanisms of crop at different growth stages, we propose a new method for extracting vegetation descriptors from Sentinel-1 dual-polarimetric SAR data. The method combines the backscattering intensity and polarization decomposition information to construct a normalized index q , which is used to generate three vegetation descriptors: the co-pol purity parameter (m cp ), the pseudo-scattering angle (θ cp ), and the pseudo-scattering entropy (H cp ). Further, a novel unsupervised clustering framework, founded on H cp and θ cp , has been proposed. This framework establishes six zones (named as Z1 to Z6) representing distinct physical scattering mechanisms, and by statistically sampling point data, it can determine the growth stage of the crops For validating the performance of the proposed vegetation descriptors and clustering framework, we conducted a three-year experiment using four crops from two publicly available datasets, namely wheat and canola from the Carman in Canada (Test site-1), corn and soybeans from Iowa in the United States (Test site-2). The experimental results indicate that m cp , θ cp , and H cp exhibit regular changes at different growth stages of crops from planting to maturity, with m cp and θ cp gradually decreasing while H cp gradually increasing. Within the entire phenology window, θ cp changes by approximately 42°, while both m cp and θ cp varies by about 0.9, and the sampling points shift from the Z2 to the Z5 zone. The vegetation descriptors are highly sensitive to the growth status of crops, and the clustering framework can also effectively respond to different growth stages of vegetation. Furthermore, the vegetation descriptors and clustering framework proposed in this study have the potential for extended application to different crop types and other polarimetric SAR data sources. [ABSTRACT FROM AUTHOR]