Spaceborne Synthetic Aperture Radar Data Research Articles

Proposal-Copula-Based Fusion of Spaceborne and Airborne SAR Images for Ship Target Detection⁎⁎

In this paper, we consider the problem of fusion of synthetic aperture radar (SAR) images from spaceborne and airborne sensors and investigate its applications to inshore ship target detection. Existing SAR image fusion methods mainly focus on image denoising or texture enhancement, but show limited improvement of target-to-clutter ratios (TCRs) in composite images and lead to deteriorated target detection performance. To address this issue, we propose a new method for the fusion of spaceborne and airborne SAR images based on the target proposal and the copula theory (TPCT). In TPCT, target and clutter correspondence between different images are exploited to improve the TCRs of composite images. TPCT consists of three steps. First, target proposals are extracted from spaceborne and airborne SAR images and then fused to enhance the common ship target areas therein. Second, a new method to construct the joint probability density function (PDF) of clutter in spaceborne and airborne SAR images is presented to model the statistical dependence of clutter therein based on the copula theory. This copula-based joint PDF is used to suppress the clutter areas remained in the intersection of target proposals. Third, clues from the intersection of target proposals and the copula-based joint PDF of clutter are fused by the Hadamard product to generate the composite image with enhanced ship targets and the suppressed clutter. Experimental results based on measured spaceborne and airborne SAR data show that the proposed TPCT fusion method leads to higher TCRs of composite images and better performance in the ship detection task than other commonly used image fusion methods.

Application of SAR Data for Tropical Cyclone Intensity Parameters Retrieval and Symmetric Wind Field Model Development

The spaceborne synthetic aperture radar (SAR) is an effective tool to observe tropical cyclone (TC) wind fields at very high spatial resolutions. TC wind speeds can be retrieved from cross-polarization signals without wind direction inputs. This paper proposed methodologies to retrieve TC intensity parameters; for example, surface maximum wind speed, TC fullness (TCF) and central surface pressure from the European Space Agency Sentinel-1 Extra Wide swath mode cross-polarization data. First, the MS1A geophysical model function was modified from 6 to 69 m/s, based on three TC samples’ SAR images and the collocated National Oceanic and Atmospheric Administration stepped frequency microwave radiometer wind speed measurements. Second, we retrieved the wind fields and maximum wind speeds of 42 TC samples up to category 5 acquired in the last five years, using the modified MS1A model. Third, the TCF values and central surface pressures were calculated from the 1-km wind retrievals, according to the radial curve fitting of wind speeds and two hurricane wind-pressure models. Three intensity parameters were found to be dependent upon each other. Compared with the best-track data, the averaged bias, correlation coefficient (Cor) and root mean-square error (RMSE) of the SAR-retrieved maximum wind speeds were –3.91 m/s, 0.88 and 7.99 m/s respectively, showing a better result than the retrievals before modification. For central pressure, the averaged bias, Cor and RMSE were 1.17 mb, 0.77 and 21.29 mb and respectively, indicating the accuracy of the proposed methodology for pressure retrieval. Finally, a new symmetric TC wind field model was developed with the fitting function of the TCF values and maximum wind speeds, radial wind curve and the Rankine Vortex model. By this model, TC wind field can be simulated just using the maximum wind speed and the radius of maximum wind speed. Compared with wind retrievals, averaged absolute bias and averaged RMSE of all samples’ wind fields simulated by the new model were smaller than those of the Rankine Vortex model.

PolSAR Data-Based Land Cover Classification Using Dual-Channel Watershed Region-Merging Segmentation and Bagging-ELM

Land cover classification based on spaceborne synthetic aperture radar (SAR) imagery is an emerging technology with many applications. Machine learning and related ensemble models have attracted much attention to this technology. This letter presents a new object-oriented land cover classification based on Polarimetric SAR data that integrate the dual-channel (DC) Watershed region-merging segmentation algorithm and a Bagging-based ensemble extreme learning machine (Bagging-ELM) image object classifier. The advantage of the DC segmentation algorithm is verified by comparing it with the well-known simple linear iterative clustering (SLIC) superpixel algorithm. Following the DC segmentation, the Bagging-ELM is implemented to perform the classification of image objects. Besides, two other classification methods (single ELM and ensembled-tree algorithms) are used for performance comparison. The quantitative analysis with respect to ground-truth information available for the test sites shows that the proposed approach can achieve noticeable improvements in classification accuracy and outperforms the other two algorithms, showing the potential for spaceborne-SAR-based land cover classification task in real applications.

Real-Time Processing of Spaceborne SAR Data With Nonlinear Trajectory Based on Variable PRF

Spaceborne synthetic aperture radar (SAR) real-time imaging is especially important for disaster emergencies and real-time monitoring applications with highly desired real-time requirements. Therefore, the continuous improvement of real-time imaging efficiency is an important development trend. At present, traditional real-time imaging algorithms based on constant pulse repetition frequency (PRF) have low accuracy when processing spaceborne SAR data with nonlinear trajectory. For this problem, the existing methods usually introduce some complex signal processing steps, such as scaling or interpolation processing, to improve the accuracy of the real-time imaging, but this will reduce its efficiency. Therefore, this article proposes a new real-time imaging algorithm based on variable PRF for nonlinear trajectory spaceborne SAR. By introducing the variable PRF, the proposed algorithm is equivalent to complete the complex signal processing steps in the radar signal transmission stage, which can greatly improve the efficiency of real-time imaging. Simulation experiments verify the effectiveness of the algorithm.

On the Processing of Gaofen-3 Spaceborne Dual-Channel Sliding Spotlight SAR Data

Currently, in the spaceborne synthetic aperture radar (SAR) community, the sliding spotlight mode has been widely applied to achieve high-resolution imaging. However, limited by the system pulse repetition frequency (PRF), it is usually impossible to realize a large-scale coverage at the same time. The technique of azimuth multichannel alleviates the PRF restriction by receiving additional spatial samples proportional to the number of Rx channels. A system combining the technique of azimuth multichannel with sliding spotlight SAR is able to achieve ultrahigh-resolution and wide-swath imaging simultaneously. In the advanced multichannel sliding spotlight mode, some problems may occur. First, azimuth beam progressive steering leads to an enlarged Doppler bandwidth over <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$N\cdot \text {PRF}$ </tex-math></inline-formula> , and conventional azimuth multichannel reconstruction algorithms developed for the stripmap mode are, thus, not directly suitable for the sliding spotlight mode. Second, the phase characteristics of multichannel sliding spotlight SAR is different from that of multichannel stripmap SAR due to the time variation of the system Doppler centroid. Third, processing of spaceborne multichannel sliding spotlight SAR experimental data shows that the convolution function used in the full-aperture reconstruction method is closely related to the antenna phase configuration strategy of the phased array antenna system. Therefore, an improved processing frame is required for the multichannel sliding spotlight SAR, reexamining the validity of assumptions, such as the stop-go-stop approximation and that of a noncurved trajectory during the relatively long aperture time. Recently, as an exploratory study, a dual-channel sliding spotlight experiment was conducted with the Gaofen-3 (GF-3) SAR. Two experimental data sets were acquired with different PRFs. In this article, the GF-3 payload and the experiment design are introduced, and a complete multichannel sliding spotlight SAR data processing chain is provided. The first imaging results of spaceborne multichannel sliding spotlight SAR are demonstrated.

A Postmatched-Filtering Image-Domain Subspace Method for Channel Mismatch Estimation of Multiple Azimuth Channels SAR

Multiple azimuth channels (MACs) synthetic aperture radar (SAR) can theoretically achieve high azimuth resolution and wide swath (HRWS). Nevertheless, in practice, channel mismatch will lead to ghost or azimuth ambiguities, which will degrade the imaging quality. This article proposes a novel idea for estimating the channel mismatch of MACs SAR in the image domain. First, we found that the degree of freedom (DOF) of MACs signals doubles after signal reconstruction and imaging. As a result, when the channel number is not great enough, the subspace method for error estimation is unable to be implemented. To deal with this problem, we introduce a DOF compression method based on spectral filtering. This method can decrease the image-domain DOF. Finally, an image-domain subspace method is proposed to estimate the channel phase error, using the focused data and selecting the high SNR region of SAR images. The proposed method has advantages for the channel phase error estimation. Simulated space-borne MACs SAR data and real measured airborne SAR data are processed to demonstrate the effectiveness of the proposed method.

Prediction of Categorized Sea Ice Concentration From Sentinel-1 SAR Images Based on a Fully Convolutional Network

The consistent and long-term spaceborne synthetic aperture radar (SAR) missions such as Sentinel-1 (S-1) provide high-quality dual-polarized <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$C$</tex-math></inline-formula> -band images particularly suited to sea ice monitoring. SAR data are currently the primary source of information for sea ice charting made by human ice experts as a result of its integration with multiple information sources at different scales (mainly radiometers). The rise of deep learning now opens the prospect of automatic sea ice mapping. In this article, we investigate the potential of a fully convolutional network (FCN) for the automatic estimation of sea ice concentration (SIC). With input data down-sampled at 200 m and an FCN architecture (depth and receptive field) duly parameterized, our approach is to mimic the work of an analyst who considers general context and does not necessarily use the highest possible resolution but speckle-noise polluted data. A comprehensive database is generated with 1320 dual-polarized S-1 scenes collocated with ice charts produced by MET Norway. A dedicated attention is paid to seasonal representativeness to ensure the adequate performance for all sea ice types. Even if the FCN output is modeled as a categorical problem, the proposed architecture accounts for the semantic distances between SIC classes by introducing an auxiliary loss. A comparative benchmark with Ocean and Sea Ice Satellite Application Facility (OSISAF) and MET Norway SIC products is carried out, showing an overall accuracy of 78.2% for our 6-class classification approach. The FCN model is shown to be evenly robust to sea ice seasonal variability and incidence angle.

TEC and Ionospheric Height Estimation by Means of Azimuth Subaperture Analysis in Quad-Polarimetric Spaceborne SAR Data

The total electron content (TEC) is probably the most important single parameter to characterize the ionospheric state and its impact on synthetic aperture radar (SAR) data. Its accurate estimation is therefore essential when it comes to the correction and calibration of SAR data distorted by the ionosphere. In recent years, the estimation of Faraday rotation (FR), i.e., the rotation of the polarimetric plane of the transmitted and scattered EM pulses as they propagate through the ionosphere, has become a key element of TEC estimation. FR is proportional to TEC and to the line-of-sight (LOS) component of the geomagnetic field. The availability of quad-polarimetric SAR data allows the precise estimation of FR, and thus TEC. However, an accurate TEC estimation requires knowledge of the geomagnetic field that varies with height. In this sense, the knowledge of the ionospheric height becomes an important issue for a correct FR to TEC conversion. While the estimation performance of FR is well understood, our understanding of the ionospheric height estimation by means of SAR is yet not established. In this article, a new estimator that allows the simultaneous estimation of TEC and ionospheric height from FR measurement is proposed. It exploits the variation of the parallel-to-LOS component of the geomagnetic field across azimuth sublooks. The performance of the proposed estimator is demonstrated using ALOS-2/PALSAR-2 data. The achieved estimation accuracy, precision, and the prerequisite for an accurate estimation are discussed.

Classification-Aided SAR and AIS Data Fusion for Space-Based Maritime Surveillance

A wide range of research activities exploit spaceborne Synthetic Aperture Radar (SAR) and Automatic Identification System (AIS) for applications that contribute to maritime safety and security. An important requirement of SAR and AIS data fusion is accurate data association (or correlation), which is the process of linking SAR ship detections and AIS observations considered to be of a common origin. The data association is particularly difficult in dense shipping environments, where ships detected in SAR imagery can be wrongly associated with AIS observations. This often results in an erroneous and/or inaccurate maritime picture. Therefore, a classification-aided data association technique is proposed which uses a transfer learning method to classify ship types in SAR imagery. Specifically, a ship classification model is first trained on AIS data and then transferred to make predictions on SAR ship detections. These predictions are subsequently used in the data association which uses a rank-ordered assignment technique to provide a robust match between the data. Two case studies in the UK are used to evaluate the performance of the classification-aided data association technique based on the types of SAR product used for maritime surveillance: wide-area and large-scale data association in the English Channel and focused data association in the Solent. Results show a high level of correspondence between the data that is robust to dense shipping or high traffic, and the confidence in the data association is improved when using class (i.e., ship type) information.

Local Katabatic Winds of the Russian Federation and Their Observation Using Synthetic Aperture Radar Imagery

Data from synthetic aperture radars (SAR) provide new opportunities for observing and studying local katabatic winds (bora and foehn) over the seas in different regions of the Russian Federation and adjacent marine basins. The impact of these winds on the characteristics of small-scale wind waves leads to the formation of characteristic manifestations on the sea surface, which in turn can be seen in SAR images. Satellite SAR data, unlike subsatellite observations, make it possible to see the phenomena in coastal zones and large lakes, estimate their spatial scales, and determine a number of quantitative characteristics. The analysis of the collected SAR images shows that bora-like and foehn-like local winds have similar characteristics and cover vast areas of the inland and marginal seas of the Russian Federation. The similarity of the manifestations of local winds clearly indicates the same physical and aerohydrodynamic mechanisms of their onset and display on SAR images. It is shown that spaceborne SAR data, which are 2D patterns of the ‘frozen wind’, are extremely useful for studying and monitoring local winds over the seas, especially where the network of weather stations is extremely sparse.

Remote Sensing of Ice Conditions in the Southeastern Baltic Sea and in the Curonian Lagoon and Validation of SAR-Based Ice Thickness Products

Here we analyze ice conditions in the Southeastern Baltic (SEB) Sea and in the Curonian Lagoon (CL) using spaceborne synthetic aperture radar (SAR) data combined with in-situ measurements from coastal stations during four winter seasons between 2009–2013. As shown, the ice conditions in the SEB and in the CL are strongly varying from year to year and do not always correlate with each other. In the SEB, ice cover may form only within 5–15 km band along the coast or spread up to 100 km offshore covering almost the entire region. The mean ice season duration here is 45 days. The CL is almost fully ice-covered every year apart of its northern part subjected to sea water inflow and active shipping. The ice regime is also more stable here, however, it also possesses multiple periods of partial melting and re-freezing. In this study we also perform a validation of three SAR-based ice thickness products (Envisat ASAR 0.5-km and 1-km, and RADARSAT-2 0.5-km) produced by the Finnish Meteorological Institute versus in-situ measurements in the CL. As shown, all satellite products perform rather well for the periods of gradual ice thickness growth. When the ice thickness grows rapidly, all products underestimate the observed values by 10–20 cm (20–50%). The best results were obtained for the RADARSAT-2 ice thickness product with the highest R2 value (0.68) and the root mean square error around 8 cm. The results of the study clearly show that multi-mission SAR data are very useful for spatial and temporal analysis of the ice regime in coastal waters and semi-enclosed shallow water bodies where the number of field observations is insufficient or lacking.

An Extended Two Step Approach to High-Resolution Airborne and Spaceborne SAR Full-Aperture Processing

The processing of synthetic aperture radar (SAR) echoes collected during radar antenna beam steering, such as in the staring spotlight, sliding spotlight, and TOPS operating modes, requires additional efforts due to the limited pulse repetition frequency (PRF). Subaperture processing is mostly used in these cases, but the complexity arising from subaperture dividing and recombination is preferably avoided. The two-step approach (TSA) is an elegant solution to full-aperture processing. Nevertheless, for high-resolution airborne and spaceborne SAR processing, the TSA is still faced with a number of difficulties, which does not occur, however, in subaperture processing. For instance, the high range bandwidth induces Doppler spectrum aliasing during the spectral analysis (SPECAN) stage of the original TSA. Moreover, the Doppler spectrum obtained in TSA, when inverse Fourier transformed back to the slow time domain, might also suffer from aliasing, which precludes the estimation and correction of the unknown residual motion error or tropospheric disturbance. In this article, we extend the TSA to address these problems and present a full-aperture processing framework to precisely focus the high-resolution airborne and spaceborne SAR data. Simulation of X-band spaceborne SAR echoes with transmission signal bandwidth of 3.6 GHz and coherent integration angle of 12.5° has validated the extended TSA (ETSA). Meanwhile, experimental X-band airborne SAR data with the same range and azimuth resolution are also processed using the ETSA, where the conventional autofocusing techniques have been smoothly incorporated.

On the Use of Multipolarization Satellite SAR Data for Coastline Extraction in Harsh Coastal Environments: The Case of Solway Firth

This study deals with coastline extraction using multipolarization spaceborne synthetic aperture radar (SAR) imagery acquired over coastal intertidal areas. The latter are very challenging environments where mud flats lead to a large variability of normalized radar cross section, which may trigger a significant number of false edges during the extraction process. The performance of SAR-based coastline extraction methods that rely on a joint combination of multipolarization information (either single- or dual-polarization metrics) and speckle filtering (either local and nonlocal approaches) are analyzed using global positioning system (GPS) samples and colocated SAR imagery collected under different incidence angles. Our test site is an intertidal zone with a wetland (i.e., salt marsh) in the Solway Firth, south-west along the Scottish-English border. Experimental results, obtained processing a pair of RadarSAT-2 full-polarimetric and a pair of Sentinel-1 dual-polarimetric SAR imagery augmented by colocated GPS samples, show that: first, the multipolarization information outperforms the single-polarization counterpart in terms of extraction accuracy; second, among the single-polarization channels, the cross-polarized one performs best; third, both single- and dual-polarization methods perform better when nonlocal speckle filtering is applied; fourth, the joint combination of nonlocal speckle filter and dual-polarization information provides the best accuracy; and finally, the incidence angle plays a role in the extraction accuracy with larger incidence angles resulting in the best performance when dual-polarization metric is used.

Advanced high-order nonlinear chirp scaling algorithm for high-resolution wide-swath spaceborne SAR

Spaceborne Synthetic Aperture Radar (SAR) is a well-established and powerful imaging technology that can provide high-resolution images of the Earth’s surface on a global scale. For future SAR systems, one of the key capabilities is to acquire images with both high-resolution and wide-swath. In parallel to the evolution of SAR sensors, more precise range models, and effective imaging algorithms are required. Due to the significant azimuth-variance of the echo signal in High-Resolution Wide-Swath (HRWS) SAR, two challenges have been faced in conventional imaging algorithms. The first challenge is constructing a precise range model of the whole scene and the second one is to develop an effective imaging algorithm since existing ones fail to process high-resolution and wide azimuth swath SAR data effectively. In this paper, an Advanced High-order Nonlinear Chirp Scaling (A-HNLCS) algorithm for HRWS SAR is proposed. First, a novel Second-Order Equivalent Squint Range Model (SOESRM) is developed to describe the range history of the whole scene, by introducing a quadratic curve to fit the deviation of the azimuth FM rate. Second, a corresponding algorithm is derived, where the azimuth-variance of the echo signal is solved by azimuth equalizing processing and accurate focusing is achieved through a high-order nonlinear chirp scaling algorithm. As a result, the whole scene can be accurately focused through one single imaging processing. Simulations are provided to validate the proposed range model and imaging algorithm.

Spaceborne L-Band Synthetic Aperture Radar Data for Geoscientific Analyses in Coastal Land Applications: A Review

The coastal zone offers among the world’s most productive and valuable ecosystems and is experiencing increasing pressure from anthropogenic impacts: human settlements, agriculture, aquaculture, trade, industrial activities, oil and gas exploitation and tourism. Earth observation has great capability to deliver valuable data at the local, regional and global scales and can support the assessment and monitoring of land- and water-related applications in coastal zones. Compared to optical satellites, cloud-cover does not limit the timeliness of data acquisition with spaceborne Synthetic Aperture Radar (SAR) sensors, which have all-weather, day and night capabilities. Hence, active radar systems demonstrate great potential for continuous mapping and monitoring of coastal regions, particularly in cloud-prone tropical and sub-tropical climates. The canopy penetration capability with long radar wavelength enables L-band SAR data to be used for coastal terrestrial environments and has been widely applied and investigated for the following geoscientific topics: mapping and monitoring of flooded vegetation and inundated areas; the retrieval of aboveground biomass; and the estimation of soil moisture. Human activities, global population growth, urban sprawl and climate change-induced impacts are leading to increased pressure on coastal ecosystems causing land degradation, deforestation and land use change. This review presents a comprehensive overview of existing research articles that apply spaceborne L-band SAR data for geoscientific analyses that are relevant for coastal land applications.

Automatic Urban Area Extraction From SAR Image Based on Morphological Operator

With the development of high-resolution spaceborne synthetic aperture radar (SAR) systems, it has been possible to use SAR data for identifying man-made structures in urban areas. In this letter, a new approach to extract urban areas in the SAR image is presented by using the morphological characteristics of the radar scattering mechanism. In order to detect the shape and location of a building, the radar shadow index and the double-bounce index are proposed. In addition, an automatic urban area extraction method is proposed using the fuzzy-based fusion of the morphological indices. The detection performance is evaluated by two ${X}$ -band SAR data acquired from different systems. Despite the differences in observation and environmental conditions in the two data, the proposed method can provide consistent detection performance with the overall accuracy of about 88% and the Kappa coefficient of about 68%.

The Potentiality of Operational Mapping of Oil Pollution in the Mediterranean Sea near the Entrance of the Suez Canal Using Sentinel-1 SAR Data

The Suez Canal, being a main international maritime shipping route, experiences heavy ship traffic with probable illegal oil discharges. Oil pollution is harming the marine ecosystem and creates pressure on the coastal socio-economic activities particularly at Port Said city (the area of study). It is anticipated that the damage of oil spills is not only during the event but it extends for a long time and normally requires more effort to remediate and recover the environment. Hence, early detection and volume estimation of these spills is the first and most important step for a successful clean-up operation. This study is the first to use Sentinel-1 space-borne Synthetic Aperture Radar (SAR) images for oil spill detection and mapping over the north entrance of the Suez Canal aiming to enable operational monitoring. SAR sensors are able to capture images day and night and are not affected by weather conditions. In addition, they have a wide swath that covers large geographical areas for possible oil spills. The present study examines a large amount of data (800 scenes of sentinel 1) for the study area over a period of five years from 2014 till 2019 which resulted in the detection of more than 20 events of oil pollution. The detection model is based on the quantitative analysis of the dark spot of the radar backscatter of oil spills. The largest case covered nearly 26 km2 of seawater. The spill drift direction in the area of spills indicated potential hazard on fishing activities, Port Said beaches and ports. This study can be the base for continuously monitoring and alarming pollution cases in the Canal area which is important for environmental agencies, decision-makers, and beneficiaries for coastal and marine socio-economic sustainability.

Validating GEV Model for Reflection Symmetry-Based Ocean Ship Detection with Gaofen-3 Dual-Polarimetric Data

The spaceborne synthetic aperture radar (SAR) is quite powerful in worldwide ocean observation, especially for ship monitoring, as a hot topic in ocean surveillance. The launched Gaofen-3 (GF3) satellite of China can provide C-band and multi-polarization SAR data, and one of its scientific applications is ocean ship detection. Compared with the single polarization system, polarimetric systems can be used for more effective ship detection. In this paper, a generalized extreme value (GEV)-based constant false alarm rate (CFAR) detector is proposed for ship detection in the ocean by using the reflection symmetry metric of dual-polarization. The reflection symmetry property shows big differences between the metallic targets at sea and the sea surface. In addition, the GEV statistical model is employed for reflection symmetry statistical distribution, which fits the reflection symmetry probability density function (pdf) well. Five dual-polarimetric GF3 stripmap ocean data sets are introduced in the paper, to show the contrast in enhancement by using reflection symmetry and to investigate the GEV model fit to the reflection symmetry metric. Additionally, with the detection experiments on the real GF3 datasets, the effectiveness and efficiency of the GEV model for reflection symmetry and the model-based ocean ship detector are verified.

A Deep Convolutional Neural Network for Oil Spill Detection from Spaceborne SAR Images

Classification algorithms for automatically detecting sea surface oil spills from spaceborne Synthetic Aperture Radars (SARs) can usually be regarded as part of a three-step processing framework, which briefly includes image segmentation, feature extraction, and target classification. A Deep Convolutional Neural Network (DCNN), named the Oil Spill Convolutional Network (OSCNet), is proposed in this paper for SAR oil spill detection, which can do the latter two steps of the three-step processing framework. Based on VGG-16, the OSCNet is obtained by designing the architecture and adjusting hyperparameters with the data set of SAR dark patches. With the help of the big data set containing more than 20,000 SAR dark patches and data augmentation, the OSCNet can have as many as 12 weight layers. It is a relatively deep Deep Learning (DL) network for SAR oil spill detection. It is shown by the experiments based on the same data set that the classification performance of OSCNet has been significantly improved compared to that of traditional machine learning (ML). The accuracy, recall, and precision are improved from 92.50%, 81.40%, and 80.95% to 94.01%, 83.51%, and 85.70%, respectively. An important reason for this improvement is that the distinguishability of the features learned by OSCNet itself from the data set is significantly higher than that of the hand-crafted features needed by traditional ML algorithms. In addition, experiments show that data augmentation plays an important role in avoiding over-fitting and hence improves the classification performance. OSCNet has also been compared with other DL classifiers for SAR oil spill detection. Due to the huge differences in the data sets, only their similarities and differences are discussed at the principle level.

Assimilation of SAR Ice and Open Water Retrievals in Environment and Climate Change Canada Regional Ice-Ocean Prediction System

In this article, we evaluate the impact of assimilating spaceborne synthetic aperture radar (SAR) data in an Arctic regional ice analysis system over a year cycle. Ice and water information was automatically extracted from more than 7000 RADARSAT-2 HH-HV ScanSAR Wide images acquired over the Canadian Arctic and adjacent waters throughout the entire year 2013. A quality-control procedure was specifically developed and applied to reduce the number of erroneous SAR retrievals. To assess the impact of SAR ice and water retrievals on the Environment and Climate Change Canada (ECCC) Regional Ice-Ocean Prediction System (RIOPS) ice concentration analyses, we designed a set of data assimilation experiments with and without the inclusion of SAR retrievals. Our verification results suggest that the assimilation of SAR-derived retrievals considerably improves ice concentration analyses in the situations where high spatial resolution is important (e.g., near land and over small inland lakes). Furthermore, SAR retrievals are particularly useful over the areas where the Canadian Ice Service's (CIS) manually derived ice products (such as Image Analyses, daily and weekly ice charts) are not available or have limited coverage. The three-satellite RADARSAT Constellation Mission (RCM) launched in June 2019 will significantly increase the temporal frequency of SAR data. According to the most recent CIS estimate, more than 54 000 RCM images a year will be acquired over the CIS areas of interest. Therefore, the assimilation of SAR retrievals from RCM should further enhance automated ice concentration analyses products.