Search

The need to monitor ice conditions has motivated the launch of several earth observation (EO) satellites and ice mapping applications are among the highest consumers of satellite data. However, oil and gas operations (O&G) in ice-prone (both sea ice and iceberg) environments have largely been using EO data for upstream, strategic reports on ice conditions. There are many recent and upcoming advances in EO technology that are already enabling satellites to be used for other critical operations and there is value in using satellites extensively for ice management. The remainder of this paper briefly describes these advances and their impact on detecting ice conditions to support oil and gas operations.

The inland water bodies are critical components of ecosystems and hydrologic cycles. Thus, the water extent data are crucially important for hydrological and ecological studies. Due to its high temporal resolution, the Cyclone Global Navigation Satellite System (CYGNSS) has the potential for real-time inland water monitoring. In this letter, a high-resolution machine learning (ML) method for detecting inland water content using the CYGNSS data is implemented via the random undersampling boosted (RUSBoost) algorithm. The CYGNSS data of the year 2018 over the Congo and Amazon basins are gridded into 0.01° x 0.01° cells. The RUSBoost-based classifier is trained and tested with the CYGNSS data over the Congo basin. The data of the Amazon basin that is unknown to the classifier are then used for further evaluation. By only using the observables extracted from the CYGNSS data, the proposed technique is able to detect 95.4% and 93.3% of the water bodies over the Congo and Amazon basins, respectively. The performance of the RUSBoost-based classifier is also compared with an image processing-based inland water detection method. For the Congo and Amazon basins, the RUSBoost-based classifier has a 3.9 % and 14.2 % higher water detection accuracy, respectively.

This letter presents the validation of an electromagnetic (EM) backscatter model of icebergs at C-band by comparing the performances of target classifiers trained with both modeled and real synthetic aperture radar (SAR) data. Simulated SAR data were obtained in a combination of imaging beam modes and scene parameters to produce 216 simulated Sentinel-1 C-band SAR images. Parameters consisted of Sentinel-1 IW1 (33.1°) and IW3 (43.1°) beam modes with varying wind speed (5 and 10 m/s), wind direction (0°, 45°, and 90°), and target orientation (0°, 45°, and 90°). Simulations were created from an EM SAR simulator called GRECOSAR, which took 3-D profiles of iceberg and ship targets and parameters necessary to closely mimic the real scenes. 3-D models of three icebergs were captured in a field study off the coast of Bonavista, Newfoundland, and Labrador, Canada in June 2017. Three generic ship models were sourced from an online inventory and scaled to a size equivalent to that of the iceberg targets. Real SAR image data were drawn from in-house data set collected from a complementary research program. Classifiers including support vector machine (SVM), Random Forest (RanFor), k-nearest neighbor (kNN), and neural network (NN) were trained with targets from modeled SAR data and then gradually mixed with real SAR data. Target classifier performance from the modeled target data was shown to be similar to classifiers trained entirely from real SAR data. The similarity in accuracy provides an indication of the validity of the modeled SAR data for this specific application.

Flash floods can cause massive damages because of their rapid evolution. To reduce or prevent harm caused by a flash flood, it is vital to have information about its formation and spread. Hence, providing real-time surveillance flood is essential. Considering Hurricane Harvey and Hurricane Irma as two case studies, six different data preparation approaches (DPAs) for flood detection based on the Cyclone Global Navigation Satellite System (CYGNSS) data and the Random Under-Sampling Boosted (RUSBoost) classification algorithm are investigated in this article. Taking flood and land as two classes, flash flood detection is tackled as a binary classification problem. Eleven observables are extracted from the delay-Doppler maps (DDMs) for feature selection. These observables, alongside two features from an ancillary data, are considered in feature selection. All the combinations of these observables with and without ancillary data are fed into the classifier with 5-fold cross-validation one by one. Based on the test results, five observables with the ancillary data are selected as a suitable feature vector for flood detection here. Using the selected feature vector, six different DPAs are investigated and compared to find the best one for flash flood detection. Then, the performance of the proposed method is compared with that of a support vector machine (SVM) based classifier. For Hurricane Harvey and Hurricane Irma, the selected method is able to detect 89.00% and 85.00% of flooded points, respectively, with a resolution of $500 \, \mathrm {m} \times 500 \, \mathrm {m}$ , and the detection accuracy for non-flooded land points is 97.20% and 71.00%, respectively.

This article presents an electromagnetic backscatter model of iceberg and compares the modeled scattering behavior with C -band RADARSAT-2 synthetic aperture radar (SAR) images. It also explores iceberg SAR signature variability over various ocean parameters. Three-dimensional (3-D) profiles of icebergs were captured in a field study off the coast of Bonavista, NL, Canada, in June 2017 at the time of an SAR satellite overpass. The 3-D profiles were captured from a vessel, using a LiDAR and multibeam sonar. The SAR image and 3-D profiles were captured within hours of one another. Simulated SAR images of the icebergs were generated in a simulator called GRECOSAR with the satellite, target orientation, and ocean parameters that closely mimic the real SAR scene. A new ocean model was introduced to model an ocean backscatter at satellites’ lower incidence angle beam mode. Comparison between real and simulated SAR images of the icebergs shows good agreement in terms of SAR signature, total radar cross section, and polarimetric decomposition. Wind direction was varied over 90° extent to observe icebergs’ backscatter variability in the simulator. Furthermore, simulated SAR images were generated for low and high wind conditions. Our study finds that the macrostructure of the melt iceberg dominates its polarimetric behavior of its backscatter. Large variability of iceberg SAR signature over varying ocean parameters was also observed. A mathematical model that considers the melting condition of iceberg suggested that significant backscattering can reflect from top surface when the melt water layer could be as little as 0.1 mm.

Ship detection and classification in very high resolution (VHR) EO/IR satellite imagery, as primary objectives, were investigated using multiple techniques. Automated algorithms were developed and their performance was evaluated using different satellite image sources (Pleiades, WorldView-2/3). Performance of ship detection algorithms based on traditional (thresholding and saliency) techniques reached probability of detection 80% for low false alarm rates. Deep learning techniques based on convolutional neural networks (CNNs) (YOLOv4 and Mask R-CNN) achieved average precision of 94–95% with 3% of false positives without the need of accurate land and cloud masking. Mask R-CNN also allows accurate determining ship size parameters. The problem of ship and non-ship classification was investigated using traditional and CNN based techniques. Linear Discriminant Analysis, Support Vector Machines and combined classifiers achieved classification accuracies close to 80–90%. At the same time, the usage of a technique based on GoogleNet CNN achieved 99% classification accuracy for ship, small boats and background targets.

The 40th Canadian Symposium on Remote Sensing was held from June 4–6, 2019, in Fredericton, New Brunswick, Canada. The Symposium was a joint event of the Canadian Remote Sensing Society and Geomati...

In this paper, the advantages of using an interferometric method for detecting and characterizing icebergs in sea ice were demonstrated. Iceberg topography was analyzed using single-pass TanDEM-X interferometric synthetic aperture radar (InSAR) data. Multiple InSAR data sets in bistatic mode were acquired over icebergs in sea ice in the Arctic region. InSAR processing was used to extract 3-D elevation information of the sea ice surface. The results firmly demonstrate the capability of TanDEM-X data to characterize the shape of icebergs. Very high resolution (VHR) optical satellite data were collected by Pleiades 1A over the same area to derive digital elevation models (DEMs) of ice features for validation. The accuracy of the extracted topography over icebergs was evaluated by comparing InSAR and optical DEMs. The quantitative comparison demonstrated good correspondence between InSAR and optical DEMs with root-mean-square value values of 2.2 m for icebergs and 0.6 m for sea ice, respectively. Using DEMs derived from VHR optical imagery, it was possible to calculate receiver operating characteristics (ROC) for detecting icebergs in sea using InSAR. The resulting ROC analysis illustrates a good detection performance.

A multi-category numerical sea ice model CICE was used along with data assimilation to derive sea ice parameters in the region of Baffin Bay and Labrador Sea. The assimilation of ice concentration was performed using the data derived from the Advanced Microwave Scanning Radiometer (AMSR-E and AMSR2). The model uses a mixed-layer slab ocean parameterization to compute the sea surface temperature (SST) and thereby to compute the freezing and melting potential of ice. The data from Advanced Very High Resolution Radiometer (AVHRR-only optimum interpolation analysis) were used to assimilate SST. The modelled ice parameters including concentration, ice thickness, freeboard and keel depth were compared with parameters estimated from remote-sensing data. The ice thickness estimated from the model was compared with the measurements derived from Soil Moisture Ocean Salinity – Microwave Imaging Radiometer using Aperture Synthesis (SMOS–MIRAS). The model freeboard estimates were compared with the freeboard measurements derived from CryoSat2. The ice concentration, thickness and freeboard estimates from the model assimilated with both ice concentration and SST were found to be within the uncertainty in the observation except during March. The model-estimated draft was compared with the measurements from an upward-looking sonar (ULS) deployed in the Labrador Sea (near Makkovik Bank). The difference between modelled draft and ULS measurements estimated from the model was found to be within 10 cm. The keel depth measurements from the ULS instruments were compared to the estimates from the model to retrieve a relationship between the ridge height and keel depth.

C-band synthetic aperture radar (SAR) is widely used for sea ice monitoring and operational activities. The RADARSAT Constellation Mission (RCM), with its anticipated launch in 2018, will provide hybrid compact polarimetric (CP) C-band SAR data offering near-polarimetric capabilities at large image acquisition widths suitable for achieving operational and scientific objectives in the Arctic. Although C-band SAR is effective for sea ice monitoring, it is difficult to implement during advanced melt, when the sea ice cover is melting and covered by melt ponds. Ice type separability during winter (pre-melt) and advanced melt conditions was assessed using Kolmogorov–Smirnov statistical separability analyses and Support Vector Machine supervised classifications of RCM parameters simulated from 2 winter and 2 advanced melt RADARSAT-2 scenes. Through a detailed analysis of the 2 advanced melt scenes, it was found that the steep incidence angle (22.3–24.2°) simulated RCM CP parameters provide improved ice ...

Use of machine learning to develop algorithms for distinguishing iceberg and vessel targets requires large validated data sets that are often costly, time consuming and, in some cases, inac...

Reductionism and holism are two worldviews underlying the fields of linear and nonlinear signal processing, respectively. Conventional radar resolution theory is motivated by the former view, and it is violated by nonlinear phase modulation induced by dispersive scattering typically associated with extended targets. Motivated by the latter view, this paper offers a new insight into the process of feature extraction for target-recognition applications in single-channel imagery output from synthetic aperture radar processors. Two novel frameworks for holism-based feature extraction are presented. The first framework is based solely on the often-ignored phase chip. The second framework uses the complex-valued 2-D synthetic aperture radar chip after it is transformed into a 1-D vector. Representative features are introduced under each framework. Further, for comparison purposes, baseline features from the power-detected chip are also considered. Three feature sets are extracted from the real-world MSTAR data set and used separately and combinatorially to design multiple instances of an eight-class support vector machine classifier. A classification accuracy of 93.42% is achieved for the holism-based features. This is in comparison to 73.63% for the baseline features. Using Fisher scoring to measure the information contained in each feature, top-ranked features from the first and second holism-based frameworks, respectively, are found to be 7 and 160 times those of the baseline features. Because the nonlinear phenomenon is resolution dependent, our proposed approach is expected to achieve even greater accuracy for synthetic aperture radar sensors with higher resolution.

The purpose of this paper is to survey and assess the state-of-the-art in automatic target recognition for synthetic aperture radar imagery (SAR-ATR). The aim is not to develop an exhaustive survey of the voluminous literature, but rather to capture in one place the various approaches for implementing the SAR-ATR system. This paper is meant to be as self-contained as possible, and it approaches the SAR-ATR problem from a holistic end-to-end perspective. A brief overview for the breadth of the SAR-ATR challenges is conducted. This is couched in terms of a single-channel SAR, and it is extendable to multi-channel SAR systems. Stages pertinent to the basic SAR-ATR system structure are defined, and the motivations of the requirements and constraints on the system constituents are addressed. For each stage in the SAR-ATR processing chain, a taxonomization methodology for surveying the numerous methods published in the open literature is proposed. Carefully selected works from the literature are presented under the taxa proposed. Novel comparisons, discussions, and comments are pinpointed throughout this paper. A two-fold benchmarking scheme for evaluating existing SAR-ATR systems and motivating new system designs is proposed. The scheme is applied to the works surveyed in this paper. Finally, a discussion is presented in which various interrelated issues, such as standard operating conditions, extended operating conditions, and target-model design, are addressed. This paper is a contribution toward fulfilling an objective of end-to-end SAR-ATR system design.

Drifting icebergs can threaten navigation and marine operations and are prevalent in a number of regions that have active oil and gas exploration and development. Satellite synthetic aperture radar (SAR) is naturally applicable to map and monitor icebergs and sea ice due its ability to capture images day or night, as well as through cloud, fog and various wind conditions. There are several notable examples of its use to support operations, including Grand Banks, Barents Sea, offshore Greenland and Kara Sea.New constellations of satellites and the increasing volume of satellite data becoming available present a new paradigm for ice surveillance, in terms of persistence, reliability and cost. To fully extract the value of the data from these constellations, automation and cloud-based processing must be implemented. This will allow more timely and efficient processing, lowering monitoring costs by at least an order of magnitude. The increase in data persistence and processing capability allows large regions to be monitored daily for ice incursions, thus increasing safety and efficiency during offshore operations in those regions.The process of automating SAR-based iceberg surveillance involves creating a process flow that is robust and requires limited human intervention. The process flow involves land-masking, target detection, target discrimination and product dissemination. Land masking involves the removal of high-clutter land from the imagery to eliminate false detection from these locations. Target detection usually involves an adaptive threshold to separate true targets from the background ocean clutter. A constant false alarm rate (CFAR) is a standard technique used in radar image processing for this purpose. Target discrimination involves an examination of the distinct features of a target to determine if they match the features of icebergs, vessels or other ‘false alarms’ (e.g., marine wildlife, clutter). The final stage is the production of an output surveillance product, which can be a standard iceberg chart (e.g., MANICE) or something that can be ingested into a GIS system (e.g., ESRI shapefile, Google KML).The target discrimination phase is one of the most important phases because it provides feedback to operations about the presence of targets of interest (icebergs and vessels). The authors have used computer vision techniques successfully to train target classifiers. Standard techniques usually result in classifier accuracies of between 85%-95%, depending on the resolution of the SAR (higher resolutions produce more accurate results) and the availability of multiple polarizations. To see if new machine learning techniques could be applied to increase classifier accuracy, a dataset of 5000 ship and iceberg targets were extracted from Sentinel-1 multi-channel data (HH,HV). The images were collected in several regions (Greenland, Grand Banks, and Strait of Gibraltar). Validation either came by way of supporting information from the offshore operations, or was inferred by location. An online machine learning competition was hosted by Kaggle, a company that conducts online competitions on behalf of their clients. The detection data were made available by Kaggle to the broad internet community. Kaggle has a loyal following of data scientists who regularly participate in Kaggle competitions. The competition was hosted over a three-month period; over 3300 teams participated in the competition. The competition produced an improved classifier over standard computer vision techniques; the top three competitors had 4-5 stage classifiers that increased classification accuracy by approximately 5%.

Target detection is the front-end stage in any automatic target recognition system for synthetic aperture radar (SAR) imagery (SAR-ATR). The efficacy of the detector directly impacts the succeeding stages in the SAR-ATR processing chain. There are numerous methods reported in the literature for implementing the detector. We offer an umbrella under which the various research activities in the field are broadly probed and taxonomized. First, a taxonomy for the various detection methods is proposed. Second, the underlying assumptions for different implementation strategies are overviewed. Third, a tabular comparison between careful selections of representative examples is introduced. Finally, a novel discussion is presented, wherein the issues covered include suitability of SAR data models, understanding the multiplicative SAR data models, and two unique perspectives on constant false alarm rate (CFAR) detection: signal processing and pattern recognition. From a signal processing perspective, CFAR is shown to be a finite impulse response band-pass filter. From a statistical pattern recognition perspective, CFAR is shown to be a suboptimal one-class classifier: a Euclidian distance classifier and a quadratic discriminant with a missing term for one-parameter and two-parameter CFAR, respectively. We make a contribution toward enabling an objective design and implementation for target detection in SAR imagery., 35 pages, 13 figures

Traditionally, the phase content in single-channel synthetic aperture radar (SAR) imagery is discarded. This practice is justified by conventional radar resolution theory, which is a theory strictly relevant to point targets. The advent of high-resolution radars permits small targets previously considered to be points to be now treated as extended targets, in which case this theory is not strictly applicable. With this in mind, this paper offers a new insight into the relevance of phase in single-channel SAR imagery. The proposed approach builds on techniques from the fields of complex-valued and directional statistics. In doing so, three main contributions are presented, the first being a novel method for characterizing the phase content. Second, a new statistical model for the phase is considered, and then a set of 15 solely phase-based features are discussed. Our results are demonstrated on real-world SAR datasets for ground-truthed targets. The statistical significance of the information carried in the phase is clearly demonstrated. Furthermore, if applied to a dataset with higher resolution, the proposed techniques are expected to achieve even higher performance.

Information on icebergs and ice islands is important for climate science and for various marine operations in Arctic and Antarctic. This work investigates capabilities of RADARSAT-2 polarimetric data for detection of icebergs in sea ice. Several iceberg detectors were analyzed with the full polarimetric data acquired in Fine Quad and Fine Quad Wide modes. The results of iceberg detection were validated with the information extracted from very high and medium resolution electro-optical satellite data including stereo datasets. It was demonstrated that the accuracy of iceberg detection depends on polarimetric bands, parameters of detection algorithm and sea ice types. The novelty of the work also includes a demonstration of detection characteristics including false alarm rates. Improvement of detection results for icebergs in pack ice was achieved using Pauli decomposition components, span and advancing detection algorithm.

Recent research has confirmed that satellite altimetry can be used for detecting icebergs. In an effort to validate the altimetry-based approach, this study used 105 samples of icebergs contained in both satellite altimeter data and ENVISAT-ASAR scenes for the Weddell Sea area. The probability of detecting icebergs larger than 150 m waterline-length was 47%. The problem of discriminating ships and icebergs based on altimeter measurements was addressed using an ensemble of automated classifiers. A total of ten features were defined from the altimetry signal to be used as predictor variables in supervised classification. The classifier ensemble comprised discriminant functions, k-nearest neighbor, neural networks, support vector machines, and decision tree analysis. Several algorithms successfully classified objects as ships or icebergs with an accuracy exceeding 85%.