MariSAR 2015: Abstracts

University College London (UCL)

Space-based SAR for maritime domain awareness

This paper argues that commercial space based synthetic aperture radar (SAR) sensors can make a significant contribution to maritime domain awareness (MDA). First the concept of MDA is defined and the motivation for it is addressed. The currently used automatic identification system (AIS) for ships is discussed at some length, and its limitations are discussed. The use of commercial space based SAR sensors for MDA is discussed in some detail. Some of the available commercial SAR satellite systems are considered in some detail. The potential of using space based EO/IR for MDA is considered as well, and the ONR Global EO MDA experiment is discussed. The conclusion is that space based SAR has a role to play in MDA.

Watch the recording

Council for Science and Industrial Research (CSIR), Pretoria, South Africa

Ocean wind, waves and current observations from SAR in South Africa

At present, very few observations of wind, waves and surface currents are available to the marine community for the coastal regions of South Africa. Satellite remote sensing observations from Synthetic Aperture Radar (SAR) can provide routine information on wave, current and winds and are used widely over the global ocean to monitor and study ocean circulation. SAR observations have remained largely unexploited in South Africa due to a lack of expertise. Here we review some of the recent research efforts in South Africa for monitoring our oceans using satellite-based SAR.

Watch the recording

Council for Science and Industrial Research (CSIR), Pretoria, South Africa

Bilge dump detection from SAR imagery using Local Binary Patterns

The detection is often done by discriminating the pollutants from the surrounding sea using Synthetic Aperture Radar (SAR) imagery. SAR is an important tool for ocean surveillance and mapping because it is one of the few sensors that can detect anomalies during the day and night through the clouds for large areas. Ocean pollutants such as bildge waste have damping effects on radar return energy, producing “darker” areas compare to the sea. This phenomena allows the detection and discrimination of sea pollutants (dark areas) from the sea cluster. Proper feature representation of the dark targets (areas containing a high probability of a bilge dump event) from the sea clutter is a crucial step as it improves performance in the discrimination of pollutants from look-alikes (dark areas formed by natural causes). A full detection system usually involves a prescreening and discrimination phase. Prescreening is the isolation of dark targets from the sea clutter. The isolated dark target from prescreening phase are processed further for separating bilge dumps from any look-alikes (this is known as the discrimination phase). The latter is highly sensitive to the performance of the prescreening phase. We investigate the use of Local Ternary Patterns derived from the popular Local Binary Patterns (LBPs) in the discrimination of possible bilge dumps from the sea cluster.

Watch the recording

University of Pisa, Italy

Spaceborne ISAR imaging for maritime surveillance

Maritime domain awareness is a challenge today because of the complexity of suh scenario and the increasing maritime traffic and number of threats. Many communication systems and sensors are employed to obtain a detailed picture of the maritime scenario at any time and in any weather condition. Because of this, radar plays a very important role in detecting, tracking and imaging targets in an area of interest. Many platforms are then used to make radar effective in a several scenarios. In particular, when large areas have to be covered, space borne radar systems are employed to obtain high resolution images in all-weather/all-time conditions. This talk will focus on spaceborne radar and will address the problem of detecting and imaging moving targets in scenarios of maritime surveillance. Specifically, the issue of focusing moving targets under complex motions (pitch, roll and yaw) will be considered and a solution based on the use of ISAR processing will be provided and validated on real Cosmo-Skymed data. A number of examples will be given to illustrate the potential of spaceborne ISAR for application of maritime surveillance.

Watch the recording

Department of Geosciences, Environment and Spatial Planning and Centro Interdisciplinar de Investigação Marinha e Ambiental (CIIMAR)

The effect of the North Equatorial Counter Current on the generation and propagation of internal solitary waves near the Amazon river shelf as observed in SAR imagery

We present a first account of Internal Solitary Waves (ISWs) on and off the Brazilian Amazon shelf, based on SAR images. The coherence crest lengths can exceed 200 km, are first detected near the southern edge of the North Equatorial Counter Current (NECC) and intensify on the northern edge of the NECC. There is a strong seasonal variability on the ISW spatial structure and propagation directions, as they enter the influence of the NECC. Some on-shelf regions are prone to intense ISW signatures in the SAR, which are believed to be associated with intricate bottom-topography.

Watch the recording

University of Porto, Portugal

Internal solitary waves in the Red Sea: SAR contribution to an unfolding mystery

The southern Red Sea is identified as a new hotspot for oceanic internal solitary waves (ISWs). SAR reveals propagation from the deep central Red Sea toward the continental shelf, where they quickly dissipate. Very weak surface tides suggest a non-trivial generation mechanism and three different possible explanations are considered: interfacial tide resonance, local generation mechanism, and locally amplified internal wave beams owing to repeated focusing reflections. Atmospheric ISWs are also briefly discussed as possible look-alikes.

Watch the recording

Florida State University, USA

Fingerprinting oil thickness on SAR

Monitoring oil spills with remote sensing is a well established task. Recently, by combining hyperspectral and microwave satellite images, we have learned that features associated with oil emulsions can be detected with SAR and optical-infrared sensors. In this talk, we present our ongoing work and our latest results, including the development of an algorithm that delineates thick oil signatures in satellite imagery. This technique is validated in a recent field campaign where oil samples were collected.

Watch the recording

Department of Ocean Sciences, Rosenstiehl School of Marine and Atmospheric Science (RSMAS)

Current measurements by TerraSAR-X and TanDEM-X along-track interferometry

For the first time, an experimental mode of TerraSAR-X permits direct surface current measurements by SAR along-track interferometry (ATI). While the system parameters are not optimal, a number of datasets acquired over river, coastal, and open ocean test sites show the potential of the technique. Inter-satellite ATI data obtained from TanDEM-X, a formation of TerraSAR-X and an identical twin, have been shown to reach a current retrieval accuracy of 0.1 m/s at an effective resolution of about 30 m. We present an overview of recent achievements and an outlook to future developments.

Watch the recording

National Oceanic and Atmospheric Administration (NOAA) Centre for Satellite Applications and Research (STAR)

Current NOAA operational SAR-derived high-resolution wind products, and plans for utilizing Sentinel-1

In May 2013, NOAA began operational production of high-resolution sea surface winds derived from RADARSAT-2 synthetic aperture radar (SAR) imagery. These unique wind products are at a resolution of 500 m and measure winds right up to the coast and in bays and straits, making these products particularly valuable for safety of coastal transportation in regions of rugged coastal topography. Current efforts to further develop this new product include use of imagery from the recently-launched Sentinel-1a satellite; new wind product formats and product distribution methods, and a product archive.

Watch the recording

Applied Physics Laboratory, University of Washington, USA

Sea-level pressure and large-scale PBL coherent structures derived from synthetic aperture radar images of the sea surface under tropical cyclones

This talk summarizes methods for retrieving Sea-level pressure from high-resolution surface wind vectors derived fromsynthetic aperture radar images of the sea surface underneath tropical cyclones. Wind vector retrieval methods use the imprint of O(500 to 3000 m) planetary boundary layer (PBL) roll vortices to establish the surface wind direction. Standard model functions are then used to derive surface wind speeds (and quality flags) from the known radar look geometry. Calibration and validation of wind vectors is generally limited to pointwise comparisons with drop sondes of other wind sources, which are of very difficult to obtain in tropical cyclone conditions. In contrast, surface pressure represents an integral measure of the scene-wind wind vector field. Furthermore, surface pressure measurements in TCs are very robust.Hence, we developed an invertible PBL model that incorporates the unique dynamics of the TC PBL from which we can derive a dynamically-constrained surface pressure pattern from the retrieved wind vectors. The same PBL model can be used to then derive an “SLP-filtered” surface wind vector field. Overall, the SLP-filtered winds are quite close to the highest-quality input wind vectors; however, they are significantly more accurate in regions of lowest-quality input wind vectors. We also investigate ubiquitous O(10 km) wavelength modulations in the surface wind convergence and surface wind stress curl. We hypothesize that they are the surface signature of very large aspect ratio PBL roll vortices and develop a nonlinear, triad wave-wave interaction theory to explain how they might be generated through upscale energy transfer from the “standard” low aspect ratio roll vortices that were used to set the surface wind direction for the wind vector retrievals.

Watch the recording

Nansen Environmental and Remote Sensing Centre, Bergen, Norway

Imaging radar satellites for coastal and ocean application with focus on the seas around South Africa

A synergetic approach for quantitative analysis of high resolution ocean imaging SAR in combination with other satellite sensors will be presented. This approach clearly demonstrates that sea surface roughness anomalies derived from these different satellite senors are spatially correlated with sharp gradients of the sea surface temperature (SST) field. To quantitatively interpret these correlated observations a direct link between the spatially varying surface roughness field and the surface divergence field is derived.

Watch the recording

University of Pretoria, South Africa

Contextual behavioural modelling and classification of vessels in a maritime piracy situation

A method is developed for modelling and classifying behaviour of maritime vessels in a piracy situation. Prior knowledge is used to construct a probabilistic graphical model of maritime vessel behaviour. This model is a novel variant of a dynamic Bayesian network (DBN), that extends the switching linear dynamic system (SLDS) to accommodate contextual information. A generative model and a classifier model are developed. The purpose of the generative model is to generate simulated data by modelling the behaviour of fishing vessels, transport vessels and pirate vessels in a maritime piracy situation. The vessels move, interact and perform various activities on a predefined map. The data generated by the model may simulate sensor data such as radar data. The model is evaluated by comparing simulation results with 2011 pirate attack reports using a novel methodology. The classifier model classifies maritime vessels into predefined categories according to their behaviour. The classification is performed by inferring the class of a vessel as a fishing, transport or pirate vessel class. The classification method is evaluated by classifying the data generated by the generative model and comparing it to the true classes of the simulated vessels.

Watch the recording

Polytechnic of Namibia/Denmark’s Technical University

Locating historic oil dumping offshore Namibia using oiled-penguin data and satellite radar imagery

Since 2002 biologists monitoring the penguin colonies near Luederitz, Namibia have recorded many episodes of finding oil-covered penguins. A likely cause is the penguins encountered oil slicks at sea from unlawful tank-discharging by ships. The events provide dates and locations to search for any existing historic SAR imagery in an attempt to find the oil slicks which the penguins encountered and get an idea of where the dumping occurred and the areal size of the slicks.

Offshore southern Namibia the large Luederitz upwelling cell of the Benguela Current brings cold, nutrient rich waters up to the surface producing a high abundance of biological activity from the base of the food chain such as plankton and algal blooms up to the largest marine animals like large populations of fish, seals and sea birds, including African Penguins. The cold current influences the weather too and foggy days are very common, especially offshore. With the hot Namib Desert coming right up to the coast and polar waters upwelling just offshore the temperature contrast is extreme, fueling coastal winds that frequently peak well over 15 m/s.

This environment makes the accurate interpretation of radar imagery challenging amongst upwelling filaments, convergent water fronts, highly variable winds, biogenic oils and there may be sub-sea oil seeps in the region too. There are also many ship wrecks on the sea floor and beach, some vessels quite large. Due to the prevalence of offshore fog using optical imagery such as from MODIS to augment the radar interpretation can not be relied upon. The search for anthropogenic oil slicks here is an ambitious task.

The Namibian coastline is uninhabited except for a handful of small communities entirely in the south. There is essentially no regular coastal observation worth mentioning and this makes it the perfect place for illegal oil dumping and according to the penguin data the ships know this. This project is a counter-measure to the oil dumping and hopes to demonstrate the utility of radar observation for environmental protection.

The salient theme of this study is to characterize the historic oil spills. Knowing the position of an oil slick tells biologists where the penguins are going in their hunting and migration routes. Physical scientists learn marine circulation patterns by comparing where slicks at sea made landfall or met a penguin from a certain colony. Matching oil slicks with any existing ship position data might also identify culprits. If the possibility of being caught by radar is publicized it might also deter ship captains from practicing illegal oil dumping.

Watch the recording

Council for Science and Industrial Research (CSIR), Pretoria, South Africa

Ship detection methods in South African waters in synthetic aperture radar imagery

The latest advances in ship detection methods in SAR imagery over South African waters with a focus on approaches using Image Processing and Pattern Recognition techniques.

Watch the recording

Centre for Maritime Research and Experimentation (CMRE)

Operational use of space based information for enhancing maritime surveillance capabilities

Sensors on Earth Observation (EO) satellites can enhance maritime surveillance capabilities, especially when treated as part of a larger sensor network and integrated with land and sea-based sensors.

Specifically, Earth Observation Synthetic Aperture Radars (EO SAR) can augment the many cooperative, transponder-based sensors with a non-cooperative capability.

This means that EO sensors have the potential to detect vessels, which are not reporting their positional information or confirm ship detections.

Watch the recording (Part 1)

Within this framework, the presentation will discuss the challenges and state of the art solutions for integrating space-based information into surveillance capabilities and will assess the advantages deriving from exploiting historical patterns of life.

The proposed methodology can be beneficial in areas with sparse data in order to propagate ship detections along confirmed routes and potentially inform EO SAR acquisition by more accurately predicting the future vessel location.

Watch the recording (Part 2)

Council for Science and Industrial Research (CSIR), Pretoria, South Africa

Use of Sentinel-1 data for vessel detection in South Africa: Early results

This talk will showcase the utility of Synthetic Aperture Radar (SAR) data acquired from the SAR instrument on-board the the recently launched sentinel-1 satellite for the detection of vessels within the South African EEZ. A number of Extra Wide (EW) Swath images were used and vessel detection methods such as standard cell averaging constant false alarm rate (CA-CFAR) were tested for various polarization modes. Early results indicate that the sentinel-1 EW images showed very promising results for vessel detection in the South African EEZ.

Watch the recording

Dept of Electrical Engineering, State University of New York at Buffalo, Amherst, New York

SAR moving target detection, geolocation and tracking

In this presentation, we outline analytical and realistic Gotcha-SAR data results on the development and merits of three SAR-MTI algorithms.

Our presentation will also include further development and refinement of our geolocation algorithms using the radial-range and angular Doppler subaperture data of a moving target.

Watch the recording

Joint Research Centre (JRC), European Commission

Maritime surveillance support: Ship classification from satellite imagery

The ocean surveillance and protection requires major efforts and an extensive sensor network, capable not only to survey coastal areas but also to monitor oceans. Considering the extent and ubiquity required for maritime surveillance, satellite imagery provides a valuable source of information. In the European Commission’s Joint Research Centre (JRC), the research focuses on building up the maritime situational picture, providing information on the active elements, collaborative and non-collaborative, by detecting, representing and classifying ships from satellite imagery.

Watch the recording

Joint Research Centre (JRC), European Commission

Maritime knowledge discovery and activity recognition from historical positioning data

Maritime Situational Awareness (MSA), the capability of understanding events, circumstances and activities within and impacting the maritime environment, can be greatly improved by the automatic identification and classification of vessel activity. In the European Commission – Joint Research Centre (JRC), the research focuses on expanding the maritime knowledge from the analysis of historical positioning data. Two research lines are pursued (1) learning maritime transport patterns to build a geographical maritime transport network and (2) maritime activity modelling and recognition.

Watch the recording

Joint Research Centre (JRC), European Commission

Using message brokering and data mediation to leverage oceanography data to help generate auxiliary information for maritime security off the Coast of Africa

The European Commission’s Joint Research Centre (JRC) has developed the Blue Hub as an in-house R&D platform for integrated maritime situational awareness. To improve maritime awareness and risk assessment, the JRC has started to integrate data from the marine science community. The National Oceanic and Atmospheric Administration is making global oceanography data available through data brokering technology. In this presentation we show how such data is being integrated into the Blue Hub system to generate auxiliary information to improve maritime situational awareness.

Watch the recording