Sea Ice Classification Research Articles

Assessing the Parameterization of RADARSAT-2 Dual-polarized ScanSAR Scenes on the Accuracy of a Convolutional Neural Network for Sea Ice Classification: Case Study over Coronation Gulf, Canada

Arctic amplification has many impacts on sea-ice extent, thickness, and flux. It becomes critical to monitor sea-ice conditions at finer spatio-temporal resolution. We used a simple convolutional neural network (CNN) on the RADARSAT-2 dual-polarized ScanSAR wide archive available over Coronation Gulf, Canada, to assess which SAR parameter improves model performances to classify sea ice from water on a large volume of data covering 11 years of ice and surface water conditions. An overall accuracy of 90.1% was achieved on 989 scenes of 100% ice cover or ice-free conditions. An accuracy of 86.3% was achieved on the last year of data (134 scenes) which was kept out of the training process to test the model on an independent dataset. A better accuracy is obtained at lower incidence angles and the HH polarization provides the most information to classify ice from water. To achieve the best accuracy, the incidence angle and the noise equivalent sigma-nought had to be included as input to the model. A comparison done with the ASI passive microwave product shows similar errors in total sea ice concentration when using the Canadian Ice Service regional charts as reference. Nonetheless, errors from both datasets differ and the CNN outputs show greater potential to reduce masked areas, given the better spatial resolution, enabling data classification closer to land and identify features not captured by the ASI dataset.

Sea Ice Classification Using Mutually Guided Contexts

In this paper, sea ice classification on a remote sensing image given just a small number of labeled pixels is investigated. Effective sea ice classification is rendered from two aspects. First, a feature extraction method is developed. It extracts the context feature from a classification map. Second, an iterative learning paradigm is established. The labeled pixels are divided into two training subsets. At each iteration, the context feature for one subset is extracted from the classification map which is obtained subject to the other subset. Therefore, the two subsets mutually guide each other for updating the context feature in an iterative manner, which finally renders effective sea ice classification. The above paradigm is referred to as mutually guided contexts. The advantages of the new paradigm are two-fold. First, the context feature enriches the sea ice image representation in a general manner regardless of the types of raw image data. Second, the two training subsets keep providing different refined classification maps for each other such that the comprehensiveness of the context feature is recursively enhanced. Therefore, the paradigm of mutually guided contexts comprehensively characterizes the sea ice image representation for training and classification even when only small training data are available. Experiments validate the effectiveness of the mutually guided contexts for sea ice classification.

Isometric mapping algorithm based GNSS-R sea ice detection

<abstract> <p>Global navigation satellite system reflectometry (GNSS-R) is based on satellite signals' multipath interference effect and has developed as one of the important remote sensing technologies in sea ice detection. An isometric mapping (ISOMAP)-based method is proposed in this paper as a development in sea ice detection approaches. The integral delay waveforms (IDWs), selected from February to April in 2018, derived from TechDemoSat-1 (TDS-1) Delay-Doppler maps (DDMs) are applied to open water and sea ice classification. In the first, the model for extracting low-dimensional coordinates of IDWs employs the randomly selected 187666 IDW samples, which are 30% of the whole IDW dataset. Then, low-dimensional coordinates of IDWs are used to train three different classifiers of support vector machine (SVM) and gradient boosting decision tree (GBDT), linear discriminant algorithm (LDA) and K-nearest neighbors (KNN) for determining the sea ice and sea water. The remaining 437889 samples, about 70% of the whole datasets, are used to validate with the ground surface type from the National Snow and Ice Data Center (NSIDC) data provided by the National Oceanic and Atmospheric Administration (NOAA). The algorithm performance is evaluated, and the overall accuracy of SVM, GBDT, LDA and KNN are 99.44%, 85.58%, 91.88% and 98.82%, respectively. The sea ice detection methods are developed, and the accuracy of detection is improved in this paper.</p> </abstract>

Deep-Learning-Based Sea Ice Classification With Sentinel-1 and AMSR-2 Data

Deep-Learning-Based Sea Ice Classification With Sentinel-1 and AMSR-2 Data

Arctic Sea Ice and Open Water Classification From Spaceborne Fully Polarimetric Synthetic Aperture Radar

We present an approach for automatic classification of Arctic sea ice and open water from spaceborne C-band RADARSAT-2 fully polarimetric (HH, HV, VH, VV) synthetic aperture radar (SAR) imagery, based on the random forest (RF) model. The HH- and HV-polarized radar backscatter, incidence angle and optimal polarimetric features are inputs of the RF model. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">First</i> , we use a physics-based unsupervised scheme to generate well-annotated sea ice and open water samples. This scheme can alleviate labeling errors arising from visual interpretation or temporal-spatial mismatch between SAR images and operational ice charts. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Subsequently</i> , we estimate a set of characteristic polarimetric parameters and select the most representative features, using the recursive feature elimination method. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Finally</i> , the RF model is trained and validated using more than one million labelled samples. Statistical validation results show that the overall sea ice and water classification accuracy is 99.94% and the Kappa coefficient is 0.999. We find that ice–water discrimination accuracy can be improved by about 4%–10%, when optimal polarimetric features are involved in the RF model input. Moreover, the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">polarization difference</i> (PD, VV–HH) is found to be the foremost polarimetric parameter for distinguishing sea ice from open water. The proposed approach has capability of yielding satisfactory ice–water classification over the complicated marginal ice zone. High-resolution ice–water classification maps clearly show that sea ice leads and their surroundings are also well separated.

Sea Ice Type and Thickness Identification Based on Vibration Sensor Networks and Machine Learning

The collision between sea ice and ship is a serious threat to navigation safety. However, the idea of using vibration data generated from ship-ice collisions to obtain physical information about sea ice, such as its type and thickness, is a novel method. This method overcomes the problem of traditional techniques (optical cameras and satellite remote sensing) being affected by visibility, and also provides data for assessing the ice load capacity of ships. In this paper, we propose a new method for sea ice type and thickness identification based on vibration sensor networks and machine learning. Firstly, the characteristic data set of sea ice type and thickness is constructed by analyzing the six nodes ice-ship vibration signals obtained from the vibration sensor network. We then used six classification algorithms to identify three common sea ice types (broken ice, flat ice, sea/open water) and assessed the accuracy of these algorithms at each node. Then, we use Gradient Boosting Decision Tree (GBDT) regression algorithm to predict sea ice thickness and evaluate the combination of node features. The results shows that the accuracy of single-node sea ice classification is 98.71%, and the R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> for sea ice thickness regression on a single node is 0.908. After feature fusion of two nodes, the R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> for sea ice thickness regression is 0.94, with a Root Mean Square Error (RMSE) of 0.847 cm, and the accuracy is better than traditional identification methods. The results show that the sea ice type and thickness can be identified with high accuracy by analyzing the vibration signals generated by ice-ship collision. The reliability and effectiveness of the proposed method in low visibility and other environments are verified by the sailing experiments in the Bohai Sea ice area accumulated at 1020 nautical miles.

Sea ice image classification based on ResFPG network and heterogeneous data fusion

ABSTRACT Sea ice detection has played an important role in climate protection and strategic deployment. Due to the autonomous learning characteristics of deep learning, it has been gradually applied to the classification of remote sensing sea ice images. At present, deep learning models are mostly used in sea ice classification of single source remote sensing data. Due to the limitations of single source data and the information loss of deep learning model in the process of feature extraction layer by layer, it is inevitable to encounter a bottleneck in sea ice detection requiring fine classification. To solve the above problems, this paper proposes a sea ice image classification method based on ResNet16-feature pyramid networks-spatial pyramid pooling-gated fusion network (ResFPG) and heterogeneous data fusion. In the feature extraction part, the method uses the improved ResNet16 to extract the multi-level feature information of sea ice in synthetic aperture radar data and optical data, reducing information loss in the feature extraction process, then mines and fuses the low-level spatial information and high-level semantic information through the improved feature pyramid networks (FPN), and then collects and fuses the output features of different scales through the spatial pyramid pooling (SPP) network. In the feature fusion part, a gated feature-level fusion strategy is designed to further improve the overall classification accuracy by adaptively adjusting the feature contribution of two heterogeneous sources of data through the gated fusion network (GFN). In order to verify the effectiveness of this method, we use two sets of heterogeneous sea ice remote sensing data located in Hudson Bay area for experiments. The experimental results show that compared with other image classification methods, the proposed method fully excavates and integrates the multi-scale and multi-level features in heterogeneous data, effectively distinguishes the feature contribution, and achieves better classification results (97.14% and 95.85%).

Remote sensing sea ice classification based on DenseNet and heterogeneous data fusion

Most remote sensing sea ice classification methods use single-source remote sensing data, such as synthetic aperture radar (SAR) data and optical remote sensing data. SAR data contain rich sea ice texture information, but the data are relatively single, making it difficult to distinguish detailed sea ice categories. Optical data include abundant spatial-spectral information, but they are often affected by clouds, fog, and severe weather. Hence, given the limitations of single-source data, the remote sensing sea ice classification accuracy cannot be further improved. A sea ice classification method based on deep learning and multisource remote sensing data fusion is proposed utilizing an improved densely connected convolutional neural network (DenseNet) to mine and fuse the multilevel features of sea ice. According to the characteristics of SAR data and optical data, a dual-branch network structure based on an improved DenseNet is employed for feature extraction, and a squeeze-and-excitation attention mechanism is introduced to weight the fusion features to further enhance the feature weights that can effectively distinguish different types of sea ice. The fully connected network is used to perform the deep fusion of features and classify sea ice. To verify the effectiveness of the proposed method, two sets of sea ice data are utilized for sea ice classification. The experimental results show that the proposed method fully excavates and fuses the multilevel characteristics of heterogeneous data using the improved dual-branch network structure, leverages the complementary characteristics of SAR data and optical data, significantly increases the sea ice classification accuracy, and effectively improves the influence of cloud cover on the sea ice classification accuracy. Compared with the typical single-source data classification method and other heterogeneous data fusion classification methods, the proposed method achieves superior overall classification accuracy (98.49% and 98.58%).

Wintertime Emissivities of the Arctic Sea Ice Types at the AMSR2 Frequencies

The surface effective emissivities of Arctic sea ice are calculated using Advanced Microwave Scanning Radiometer 2 (AMSR2) measurements. These emissivities are analyzed for stable winter conditions during the months of January–May and November and December of 2020 for several main sea ice types defined with the sea ice maps of the Arctic and Antarctic Research Institute (AARI). The sea ice emissivities are derived from the AMSR2 data using the radiation transfer model for a non-scattering atmosphere and ERA5 reanalysis data. The emissivities are analyzed only for areas of totally consolidated sea ice of definite types. Probability distribution functions are built for the emissivities and their functions for such sea ice types as nilas, young ice, thin first-year (FY) ice, medium FY ice, thick FY ice and multi-year ice. The emissivity variations with frequency are estimated for each of the considered sea ice type for all seven months. The variations are calculated both for the emissivities and for their gradients at the AMSR2 channel frequencies. Obtained emissivities turned out to be generally lower than reported previously in scientific studies, whereas the emissivity variability values proved to be much larger than was known before. For all FY ice types, at all the frequencies, an increase in the emissivity at the beginning of winter and its decrease by the end of May are observed. The emissivity gradients demonstrate noticeable decreases with sea ice age, and their values may be used in sea ice classification algorithms based on the AMSR2 data.

C- and L-band SAR signatures of Arctic sea ice during freeze-up

Identifying sea ice types in the early stages of development from L-band SAR imagery remains an active research area during the Arctic freeze-up period. We used ScanSAR C- and L-band imagery from RADARSAT-2, ALOS PALSAR and ALOS-2 PALSAR-2, to identify ice types in the North Water Polynya (NOW) and Victoria Strait (VS) region of the Canadian Arctic. We investigated the HH-polarized microwave backscatter coefficient (σHH0) and its GLCM texture parameters for six ice classes and open water. We found very low σHH0 for nilas at both C- and L-band. Although similar σHH0 found for grey ice at both frequencies, σHH0 decrease with increasing ice thickness at L-band from grey ice, whereas, at C-band, σHH0 increases from grey to grey-white ice and then decreases as the ice grows. GLCM texture parameters show lower values for L-band than C-band; however, separability among classes was found only for a few selected parameters. We used the support vector machine (SVM) algorithm for ice type classification from SAR scenes using σHH0 and GLCM texture statistics. Due to overlapping σHH0 signatures at C-band, early-stage ice classes were substantially misclassified. L-band identified early-stage ice classes with higher accuracy compared to C-band but misclassified thicker ice types and open water. L-band alone provided very good classification results (~80% accuracy) and combining L- and C-band (i.e., dual-frequency approach) further increased accuracy to >90%. C-band alone resulted in the lowest accuracy of <60%. We acknowledge that developing a universal ice classification is still a challenge and requires some manual supervision to adopt variable ice conditions into the classification method. However, a dual-frequency approach can achieve higher classification accuracy than conventionally used single-frequency approaches. This research highlights the value of upcoming L-Band SAR missions to improve sea ice classification in regions where a variety of ice types exist, including many thinner types, which are now dominating an increasingly warming Arctic.

Sea Ice–Water Classification of RADARSAT-2 Imagery Based on Residual Neural Networks (ResNet) with Regional Pooling

Sea ice mapping plays an integral role in ship navigation and meteorological modeling in the polar regions. Numerous published studies in sea ice classification using synthetic aperture radar (SAR) have reported high classification rates. However, many of these focus on numerical results based on sample points and ignore the quality of the inferred sea ice maps. We have designed and implemented a novel SAR sea ice classification algorithm where the spatial context, obtained by the unsupervised IRGS segmentation algorithm, is integrated with texture features extracted by a residual neural network (ResNet) and, using regional pooling, classifies ice and water. This algorithm is trained and tested on a published dataset and cross-validated using leave-one-out (LOO) strategy, obtaining an overall accuracy of 99.67% and outperforming several existing algorithms. In addition, visual results show that this new method produces sea ice maps with natural ice–water boundaries and fewer ice and water errors.

Year-round sea ice and snow characterization from combined passive and active microwave observations and radiative transfer modeling

Satellite microwave observations from 1.4 to 36 GHz already showed sensitivity to several geophysical parameters of sea ice such as Sea Ice Concentration (SIC), Sea Ice Thickness (SIT) or snow depth. The main goal of this article is to provide a realistic and comprehensive characterization of the sea ice and its snow cover that explains the microwave observations during a whole year using a radiative transfer model. For this purpose, we construct a unique dataset of passive microwave observations, to mimic the future Copernicus Imaging Microwave Radiometer (CIMR), along with the active microwave scatterometer data (ASCAT). CIMR database is used to classify sea ice microwave signatures in their spectral dimension with a machine learning technique while ASCAT data are used to help interpret the results of the classification. Classification results are then interpreted with a state-of-art sea ice and Snow Microwave Radiative Transfer model (SMRT) for all highlighted signatures and all seasons. Results make it possible to identify the specific behaviors from the observation co-variabilities for SIC, SIT, and snow structure. Our analysis underlined the role of the depth hoar over multi-year ice, for the interpretation of scattering signals in winter. Scattering signals that appear in late summer are explained by the presence of superimposed ice. This characterization will benefit from future advances in SMRT development, as well as the improved observations of future satellite missions.

Eastern Arctic Sea Ice Sensing: First Results from the RADARSAT Constellation Mission Data

Sea ice monitoring plays a vital role in secure navigation and offshore activities. Synthetic aperture radar (SAR) has been widely used as an effective tool for sea ice remote sensing (e.g., ice type classification, concentration and thickness retrieval) for decades because it can collect data by day and night and in almost all weather conditions. The RADARSAT Constellation Mission (RCM) is a new Canadian SAR mission providing several new services and data, with higher spatial coverage and temporal resolution than previous Radarsat missions. As a very deep convolutional neural network, Normalizer-Free ResNet (NFNet) was proposed by DeepMind in early 2021 and achieved a new state-of-the-art accuracy on the ImageNet dataset. In this paper, the RCM data are utilized for sea ice detection and classification using NFNet for the first time. HH, HV and the cross-polarization ratio are extracted from the dual-polarized RCM data with a medium resolution (50 m) for an NFNet-F0 model. Experimental results from Eastern Arctic show that destriping in the HV channel is necessary to improve the quality of sea ice classification. A two-level random forest (RF) classification model is also applied as a conventional technique for comparisons with NFNet. The sea ice concentration estimated based on the classification result from each region was validated with the corresponding polygon of the Canadian weekly regional ice chart. The overall classification accuracy confirms the superior capacity of the NFNet model over the RF model for sea ice monitoring and the sea ice sensing capacity of RCM.

An Improved Sea Ice Classification Algorithm with Gaofen-3 Dual-Polarization SAR Data Based on Deep Convolutional Neural Networks

The distribution of sea ice is one of the major safety hazards for sea navigation. As human activities in polar regions become more frequent, monitoring and forecasting of sea ice are of great significance. In this paper, we use SAR data from the C-band synthetic aperture radar (SAR) Gaofen-3 satellite in the dual-polarization (VV, VH) fine strip II (FSII) mode of operation to study the Arctic sea ice classification in winter. SAR data we use were taken in the western Arctic Ocean from January to February 2020. We classify the sea ice into four categories, namely new ice (NI), thin first-year ice (tI), thick first-year ice (TI), and old ice (OI), by referring to the ice maps provided by the Canadian Ice Service (CIS). Then, we use the deep learning model MobileNetV3 as the backbone network, input samples of different sizes, and combine the backbone network with multiscale feature fusion methods to build a deep learning model called Multiscale MobileNet (MSMN). Dual-polarization SAR data are used to synthesize pseudocolor images and produce samples of sizes 16 × 16 × 3, 32 × 32 × 3, and 64 × 64 × 3 as input. Ultimately, MSMN can reach over 95% classification accuracy on testing SAR sea ice images. The classification results using only VV polarization or VH polarization data are tested, and it is found that using dual-polarization data could improve the classification accuracy by 10.05% and 9.35%, respectively. When other classification models are trained using the training data from this paper for comparison, the accuracy of MSMN is 4.86% and 1.84% higher on average than that of the model built using convolutional neural networks (CNNs) and ResNet18 model, respectively.

Cross-platform classification of level and deformed sea ice considering per-class incident angle dependency of backscatter intensity

Abstract. Wide-swath C-band synthetic aperture radar (SAR) has been used for sea ice classification and estimates of sea ice drift and deformation since it first became widely available in the 1990s. Here, we examine the potential to distinguish surface features created by sea ice deformation using ice type classification of SAR data. Also, we investigate the cross-platform transferability between training sets derived from Sentinel-1 Extra Wide (S1 EW) and RADARSAT-2 (RS2) ScanSAR Wide A (SCWA) and fine quad-polarimetric (FQ) data, as the same radiometrically calibrated backscatter coefficients are expected from the two C-band sensors. We use a novel sea ice classification method developed based on Arctic-wide S1 EW training, which considers per-ice-type incident angle (IA) dependency of backscatter intensity. This study focuses on the region near Fram Strait north of Svalbard to utilize expert knowledge of ice conditions during the Norwegian young sea ICE (N-ICE2015) expedition. Manually drawn polygons of different ice types for S1 EW, RS2 SCWA and RS2 FQ data are used to retrain the classifier. Different training sets yield similar classification results and IA slopes, with the exception of leads with calm open water, nilas or newly formed ice (the “leads” class). This is caused by different noise floor configurations of S1 and RS2 data, which interact differently with leads, necessitating dataset-specific retraining for this class. SAR scenes are then classified based on the classifier retrained for each dataset, with the classification scheme altered to separate level from deformed ice to enable direct comparison with independently derived sea ice deformation maps. The comparisons show that the classification of C-band SAR can be used to distinguish areas of ice divergence occupied by leads, young ice and level first-year ice (LFYI). However, it has limited capacity in delineating areas of ice deformation due to ambiguities between ice types with higher backscatter intensities. This study provides reference to future studies seeking cross-platform application of training sets so they are fully utilized, and we expect further development of the classifier and the inclusion of other SAR datasets to enable image-classification-based ice deformation detection using only satellite SAR.

Investigation of Polarimetric Decomposition for Arctic Summer Sea Ice Classification Using Gaofen-3 Fully Polarimetric SAR Data

The aim of this article was to investigate the potential of polarimetric decomposition of Chinese Gaofen-3 (GF-3) C-band fully polarimetric synthetic aperture radar (PolSAR) data for Arctic sea ice classification during summer season. Five different polarimetric decomposition approaches, including the Cloude-Pottier decomposition (Cloude), the Freeman three-component decomposition (Freeman3), the Freeman three-component decomposition using the extended Bragg model (Freeman3X), the Yamaguchi three-component decomposition (Yamaguchi3), and the nonnegative eigenvalue decomposition (NNED) were analyzed using 35 scenes of GF-3 PolSAR data collected over the Fram Strait, Arctic from June 14–18, 2017. Polarimetric features extracted from these five methods were evaluated and utilized to train random forest classifiers to classify open water (calm water and rough water) and sea ice types (melted ice, unmelted ice, and deformed ice). The results show that NNED could ensure physically valid decomposed powers while the other three model-based decompositions had negative values. In terms of sea ice classification, NNED had the highest feature importance scores and achieved an overall accuracy and Kappa coefficient of about 86.18% and 0.82, respectively. Inclusion of radar incidence angle as a feature in the classifier could slightly improve the classification accuracy by about 3%. The influence of incidence angle on sea ice classification accuracy was also investigated and it was found that high incidence angles (39°–46°) were superior to low incidence angles (21°–27°) due to the overall higher accuracies.

Intercomparison of Arctic Sea Ice Backscatter and Ice Type Classification Using Ku-Band and C-Band Scatterometers

As a result of global warming, multiyear ice (MYI) is being replaced by first-year ice (FYI) in the Arctic. Microwave scatterometers in the Ku-band and C-band can provide daily observations of sea ice type. However, their comparative capabilities in mapping ice type have not been thoroughly evaluated. We present a systematic intercomparison of the backscatter signature in VV polarization ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\sigma }_{\mathrm {vv}}^{\mathrm {o}}$ </tex-math></inline-formula> ) and the sea ice classification from three scatterometer systems using the same ice classification approach. The systems are the Ku-band quick scatterometer (QSCAT) and the newly launched Chinese rotating fan-beam scatterometer (RFSCAT) and the C-band advanced scatterometer (ASCAT). Three freezing seasons are used, i.e., 2007/08 and 2008/09 for the QSCAT/ASCAT comparison and 2019/20 for the RFSCAT/ASCAT comparison. With reference to ASCAT, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\sigma }_{\mathrm {vv}}^{\mathrm {o}}$ </tex-math></inline-formula> bias between QSCAT and RFSCAT results from their different incidence angles. A continuous declining trend of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\sigma }_{\mathrm {vv}}^{\mathrm {o}}$ </tex-math></inline-formula> from MYI and FYI is observed during winter, with a greater difference between MYI and FYI in the Ku-band. The MYI and FYI extent derived from QSCAT/RFSCAT is highly consistent with that derived from ASCAT, with a difference less than 7% and 3% for MYI and FYI, respectively. The overall accuracy (OA) is around 77% and 80% for the RFSCAT results and ASCAT results, respectively, compared with Sentinel-1 SAR images. The classification results show high consistency (81%–89%) with ice charts from the Canadian Ice Service. The incorporation of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {Tb}}_{36\mathrm {h}}$ </tex-math></inline-formula> from AMSR-E/AMSR2 improves the OA of the classification when using ASCAT or RFSCAT by 7%–11%.

Sea-Ice Mapping of RADARSAT-2 Imagery by Integrating Spatial Contexture With Textural Features

Mapping different types of sea ice that form, grow, and melt in polar oceans is essential for shipping navigation, climate change modeling, and local community safety. Currently, ice charts are manually generated by analysts at the Canadian Ice Service (CIS) based on dual-polarized RADARSAT-2/RADARSAT Constellation Mission (RCM) imagery on a daily basis. Inspired by the demand for a computer-based mapping system, we have developed an automatic sea ice classification method that integrates spatial contexture (unsupervised segmentation) with textural features (supervised pixel-level labeling). First, the full-scene image is oversegmented, and the segments are merged into homogeneous regions across the entire scene. Second, pixel-based classifiers (support vector machine, random forest) are compared for their ability to label the generated homogeneous regions. Finally, the segmentation and labeling are combined using a proposed energy function. The proposed method was tested on 18 dual-polarization RADARSAT-2 scenes acquired over the Beaufort Sea. This dataset contains water, young ice, first-year ice, and multi-year ice covering melt, summer, and freeze-up seasons. The proposed method obtains an average classification accuracy of 86.33% based on the leave-one-out validation. The experimental results show that the proposed method achieves promising classification results in both quantity and quality measurements compared to benchmark methods. The robustness against incidence angle variance indicates that the proposed method is well-qualified for operational sea ice mapping.

A Sea Ice Recognition Algorithm in Bohai Based on Random Forest

As an important maritime hub, Bohai Sea Bay provides great convenience for shipping and suffers from sea ice disasters of different severity every winter, which greatly affects the socio-economic and development of the region. Therefore, this paper uses FY-4A (a weather satellite) data to study sea ice in the Bohai Sea. After processing the data for land removal and cloud detection, it combines multi-channel threshold method and adaptive threshold algorithm to realize the recognition of Bohai Sea ice under clear sky conditions. The random forests classification algorithm is introduced in sea ice identification, which can achieve a certain effect of sea ice classification recognition under cloud cover. Under non-clear sky conditions, the results of Bohai Sea ice identification based on random forests have been improved, and the algorithm can effectively identify Bohai Sea Ice and can improve the accuracy of sea ice identification, which lays a foundation for the accuracy and stability of sea ice identification. It realizes sea ice identification in the Bohai Sea and provides data support and algorithm support for marine climate forecasting related departments.

SI-STSAR-7: A Large SAR Images Dataset with Spatial and Temporal Information for Classification of Winter Sea Ice in Hudson Bay

Remote sensing satellites have been broadly applied to sea ice monitoring. The substantial increase in satellite imagery provides a large amount of data support for deep learning methods in the sea ice classification field. However, there is a lack of public remote sensing datasets to facilitate sea ice classification with spatial and temporal information and to benchmark the deep learning methods. In this paper, we provide a labeled large sea ice dataset derived from time-series sentinel-1 SAR images, dubbed SI-STSAR-7, and a validated dataset construction method for sea ice classification research. The SI-STSAR-7 dataset includes seven different sea ice types corresponding to different sea ice development stages in Hudson Bay during winter, and its samples are time sequences of SAR image patches in order to embody the differences of backscattering intensity and textures between different sea ice types, as well as the change of sea ice with time. We construct the dataset by first performing noise reduction and mitigation of incidence angle dependence on SAR images, and then producing data samples and labeling them based on our proposed sample-producing principles and the weekly regional ice charts provided by Canadian Ice Service. Three baseline classification methods are developed on SI-STSAR-7 to establish benchmarks, which are evaluated with accuracy and kappa coefficient. The sample-producing principles are verified through experiments. Based on the experimental results, sea ice classification can be implemented well on SI-STSAR-7.