Search

Synthetic Aperture Radar (SAR) and medium (MR) and very high (VHR) resolution optical satellite images are valuable sources of information for maritime situational awareness. This paper presents the Earth Observation Maritime Surveillance System (EO-MARISS), an integrated processing framework of remote sensing satellite data developed at DLR’s Maritime Safety and Security Lab Neustrelitz to support sea monitoring tasks. The system is currently developed for operational use at DLR’s Ground Station Neustrelitz to support service provider like AIRBUS or EUSI and processes image data from different satellite missions received directly at the ground station or provided by a network of ground stations and service providers.

The focus of the study is analysing storm peak/center propagation, front movement and arrival of swell using newest remote sensing information, numerical models, in-situ measurements and their combination. For this purposes, a new empirical algorithm for sea state retrieval from satellite borne Sentinel-1 (S1) Synthetic Aperture Radar (SAR) imagery was developed. The algorithm is applied inside a new processor for meteo-marine parameter estimation for Near Real Time (NRT) applications. These NRT-applications include the investigation of geophysical processes using different satellite modes ranging from high resolution modes with small image coverage of ~20km in open ocean to low resolution modes with wide coverage of ~250km in shelf areas.The quick developments in satellite techniques, processors, algorithms and ground infrastructures provide new possibilities for series of oceanographic applications in the last years. These new techniques allow estimation of a wide range of oceanographic information including properties of surface waves and internal waves, surface wind speed, sub-meso scale fronts and eddies, ice coverage, oil spills, coastal bathymetry, currents and others. Generally, the new high resolution products from different models allow verification of meteo-marine parameters more accurately. Here, a cross validation with different sea state model results using WWIII (NOAA) and CMEMS (COPERNICUS), with in situ buoy measurements and with satellite estimated parameters allowed an significant improvement of the accuracy of the derived sea state and wind fields. The new empirical algorithm allows estimation of total integrated sea state parameters including significant wave height Hs, first moment wave period Tm1, second moment period Tm2, mean period Tm and also partial integrated parameters like swell and windsea wave heights and windsea period. The algorithm allows processing of different S1 Synthetic Aperture Radar (SAR) modes into sea state fields: S1 Wave Mode (WV) acquires multiple vignettes with an extent of ~20km×20km and each displaced by 100 km along satellite tracks in open ocean (global). About 60 tracks around the globe have been acquired per day. The relatively high spatial resolution of ~4 m allows estimating wave height with accuracy of ~35cm. This is comparable with the accuracy of satellite altimetry and a new achievement for SAR based techniques. S1 Interferometric Wide Swath Mode (IW) covers area-strips of thousand kilometres of earth and ocean surface in coastal areas with a resolution of ~30m by sequences of multiple images with an approximate size of 200km×250km. The accuracy of ~ 70cm (Hs) for this mode is not as so high as for S1-WV, because the short waves are not visible for S1-IW mode and imaged as noise. However, the accuracy is much higher than state-of-the-art methods for this mode. The algorithm has been integrated into a prototype processor for Sentinel-1 SAR imagery. The DLR Ground Station Neustrelitz applies this prototype as part of a near real-time demonstrator MSA service. The presented scientific service involves daily provision of surface wind and sea state parameters estimated fully automatically from S1 IW images of North and Baltic Sea in and around German territorial waters.

A new empirical algorithm CWAVE_S1-IW for estimation of significant wave height Hs including swell and wind sea from C-band satellite-borne Synthetic Aperture Radar (SAR) data has been developed for Sentinel-1 (S1) Interferometric Wide Swath Mode (IW) imagery. The algorithm was implemented into the Sea Sate Processor (SSP) for fully automatic processing for near real time (NRT) services and allow the estimation of wave fields of thousands of kilometres in the flight direction and 250 km swath from S1 IW scenes consisting of a sequence of individual images. The priority of CWAVE_S1-IW development was an automatic, fast and robust raster processing independent from wave patterns, applicable even when only clutter is visible in the SAR images. The algorithm is based on the spectral analysis of subscenes in wavenumber space. The empirical function allows direct Hs estimation from image spectra without first converting them into wave spectra and uses integrated image spectra parameters as well as estimated local wind information. A texture analysis based on Grey Level Co-occurrence Matrices (GLCM) is also applied. In this way, also the parameters of short waves can be estimated, which are not visible in S1 IW images and are only represented by clutter. The algorithm was tuned worldwide using in-situ collocated measurements of 92 buoys with more than 2500 acquisitions. The validated SSP allows automatic processing of worldwide S1 IW images in VV or HH polarization, including Atlantic storms, cyclones and huge storms in the Gulf of Alaska with an root mean square error RMSE of 80 cm for Hs. For the closed seas like the North Sea, Baltic Seas and Black Sea the accuracy is higher with an RMSE = 55 cm. The algorithm is integrated into a demonstration service, used for further validation at the DLR ground station in Neustrelitz. The NRT processing has been tested by supporting a research ship cruise in the Antarctic Sea.

Different metocean conditions have an impact on the detectability of ship signatures on Synthetic Aperture Radar (SAR) images. During the EMSec Project algorithms for retrieval of wind and sea state fields from TerraSAR-X data have been developed in conjunction with a near real-time-capable constant false alarm rate ship detection processor. This paper presents a new model connecting these three information extraction systems into a ship detectability model by setting the probability of detection in dependency to the four parameters: Wind speed, significant wave height, incidence angle and ship length. The model is based on a binary L2-regularized logistic regression classifier trained on a large dataset of X-band SAR ship samples, which are identified using Automatic Identification System messages co-located automatically in space and time and further checked manually to avoid possible mismatches. Results are compared to the state-of-the-art simulation-based ship detectability model available in literature. For the first time it has been possible to evaluate not only qualitatively but also quantitatively the effects of acquisition geometry and metocean conditions for the different image resolution classes obtainable with the high-flexible SAR sensor on-board the TerraSAR-X satellite.

From Synthetic Aperture Radar (SAR) imagery various types of maritime information are extractable, including surface winds, waves, oil slicks, waterline changes, changes of seabed morphology in shallow waters, wakes of ships, etc. The algorithms recently developed for this purpose are integrated into a prototype processor for Sentinel-1 (S-1) imagery. The DLR Ground Station Neustrelitz applies this prototype as part of a near real-time demonstrator service for support of Maritime Situation Awareness. The presented scientific service involves daily provision of surface wind and sea state parameters estimated fully automatic from S-1 Interferometric Wide Swath (IW) images for North and Baltic Sea.

Synthetic Aperture Radar (SAR) satellites are able to observe small and large scale structures in sea ice - in any weather, through clouds and darkness. In order to assist ship navigation during polar campaigns, we acquired SAR images along the ship course and provided them to navigators on board in near real time, utilizing the operational data processing chain of DLR ground station Neustrelitz. These “exclusive” acquisitions already helped to optimize the routes. SAR data, however, contain more information that is not easily visible, e.g. information about the local sea ice drift. In this paper, we explore the capabilities of a new software processor that is intended to retrieve high resolution sea ice drift fields from pairs of colocated SAR images, combining TerraSAR-X and Radarsat-2 images. The processor is foreseen to be integrated into the operational data processing chain at DLR ground station network sites.

The satellite of the TerraSAR-X mission, called TSX, launched on June 15th 2007 and its identically constructed twin satellite TDX, which is required by the mission TanDEM-X, launched on 21st of June 2010. Together they supply high-quality radar data in order to serve two mission goals: Scientific observation of the Earth and provision of remote sensing data for the commercial market (TerraSAR-X mission) and the generation of a global digital elevation model (DEM) of the Earth’s surface (TanDEM-X mission). On the occasion of the 10th anniversary of the mission, the focus will be on the development of the TerraSAR-X system during this period, including the extension of the ground segment, the evolution of the product portfolio, dedicated mission campaigns, radar experiments, refinement of the satellite operations and orbit control and results of the performance monitoring. Despite numerous interventions in the overall system, mainly driven by science and commercial demands in the past ten years we managed to incorporate new scientific and commercial requirements and to improve and enhance the overall system in order to fulfill the increasing demand for earth observation data without noticeable interruptions to ongoing operations.

Space-borne Synthetic Aperture Radar (SAR) images have proven to be very useful tool in challenging Polar environment - due to its active radar antenna which can acquire valuable data in all weathers, though clouds and darkness. In several polar expeditions, German Aerospace Center (DLR) acquired high-resolution SAR images from the satellite mission TerraSAR-X/TanDEM-X over the planned course, and provided them to navigators on board the ship in near real-time. It has been shown that these targeted acquisitions helped the onboard navigators to identify cost effective routes. During the course of MOSAiC campaign German Aerospace Center (DLR) plans to provide SAR derived sea ice related products such as sea ice type, sea ice dirft information within Near Real Time through AWI’s MapViewer. The sea ice type product is based on dual polarized SAR acquisitions and capable to discriminate between open water and different types of sea ice (e.g. First Year Ice, Multi Year Ice). The sea ice drift algorithm is based on two consecutive SAR acquisition can capable to provide high resolution drift information within Near Real Time. During the Potsdam workshop we will show examples of above mentioned products which is not only helpful for campaign planning but also might provide useful information to scientists across different scientific domains.

Innovative algorithms for Sentinel-1 (S1) satellite allow daily observations of meteo-marine parameters, tracking of storm propagation, study of local sea state variability and coastal processes. In order to investigate geophysical processes, the sea state and wind fields estimated simultaneously from S1 scenes acquired twice daily are combined with numerical forecast model results and in-situ measurements. The focus of the investigations are the evolution and propagation of storm peaks/centres, storm front movement, and arrival of swell. An example of efficient storm tracking in the Black Sea in April 2017 over three days is analysed. The HZG forecast spectral wave model running for the Black Sea reproduces the storm peak propagation near to the S1 observations. In detail, the storm peak observed by S1 is shifted ~80km towards the south in comparison to the model simulations. During this storm, the ship “Geroi Arsenala” of river-sea class licensed for inland waterways with access to the coastal seas was capsized about 40 kilometres to the south of the Kerch Strait in the open Black Sea, according to the associated press. The cargo ship was carrying grain from Russia to Turkey. Obviously, the course was taken too far from the coast to shorten the way across the sea with unexpected high sea state. Only one of the 12 people aboard was rescued. This is a tragic incident that proves the importance of Maritime Safety and Security. With the Sentinel satellites and the processing framework demonstrated in this work, we have appropriate tools to raise Maritime Situation Awareness (MSA) to unprecedented levels, which helps avoiding such accidents. The rapid development of satellite techniques, information extraction algorithms and ground infrastructures during the last years enabled a series of oceanographic applications with near-real-time (NRT) capabilities. Several minutes after acquisition, the produced data with geo-coded information on wind speed and wave height can be transferred to the weather services for validation of the forecasting models. The different kind of data like coastline, wave height, surface wind speed, ice coverage, oil spills etc. can be processed in parallel for the same image and combined with other information (e.g. model results, ship traffic) for supporting MSA. The algorithms currently developed for this purpose are integrated into a prototype processor for Sentinel-1 imagery. The DLR Ground Station Neustrelitz applies this prototype as part of a near real-time demonstrator MSA service. The presented scientific service involves daily provision of surface wind and sea state parameters estimated fully automatically from S1 Wide Swath Mode (IW) Synthetic Aperture Radar (SAR) images of North and Baltic Sea. S1 IW covers area-strips of thousand kilometres of earth and ocean surface with a resolution of ~10m by sequences of multiple individual IW images with an approximate size of 200km×250km. The data are free to use and provide unprecedented observation possibilities of ocean processes and natural phenomena worldwide at a high repetition rate. Due to the independence of sunlight and cloud coverage, SAR data are an indispensable source of 2D information of the ocean surface for open sea and for coastal applications.

The automatic daily processing of coastlines from Sentinel-1 satellite data for updating the bathymetry to support modelling as part of a Near Real Time chain is being developed. In shallow waters, the physical processes caused by the interaction between waves, currents and the sea bottom become important and are crucial for modelling, requiring current and accurate bathymetry data. The spatial properties of sea state and currents change strongly in coastal areas and especially in the Wadden Sea at the Danish, German and Dutch North Sea coast with large areas of intertidal flats. The morphodynamics of seabed structures are significant in such littoral zones; the soft seabed can change within days during severe storms. Data from past measurement campaigns, which are very expensive using ship soundings and airborne LIDAR scanning, deprecates quickly as the positions of islands, sandbanks or tidal inlets change. Acquisitions from the Sentinel-1 (S1) Synthetic Aperture Radar (SAR) satellites cover most coasts worldwide. With two satellites currently in orbit, the acquisition interval is at most 6 days at the equator and almost daily in our latitudes. The Interferometric Wide Swath (IW) mode covers lands and coastal waters with a swath width of 250km and an image resolution of 10m. In contrast to optical satellites, SAR acquisitions are independent of illumination and cloud cover; hence, they deliver new images very reliably. The DLR satellite ground station in Neustrelitz receives these data and processes and delivers them in Near Real Time (NRT), usually within 20 minutes after acquisition. We have developed automated algorithms to retrieve the waterline from SAR images. In the tidal flat areas of the Wadden Sea, analyzing time series of acquisitions at similar tidal states allows estimating rates of change and sediment transport; a combination of several images acquired within a short time frame at different tidal states allows estimating the topography. Our algorithms for Sentinel-1 SAR images offer a quick and cheap way to spot ongoing processes, verify modelling results and identify regions of major change for future measurement campaigns.

Maritime Situational Awareness is of global importance and satellite remote sensing technologies are well suited to derive maritime value added products. Processing systems to derive information about oil spill, ship or iceberg detection from spaceborne Synthetic Aperture Radar (SAR) image data are being developed and integrated at the Ground Station Neustrelitz, part of the German Remote Sensing Data Center (DFD). Shortly, products of meteo-marine parameters like wind and wave will complement the product portfolio. Research and development aim at the implementation of highly automated services for operational use. At this time SAR images are widely used because of the possibility to provide maritime products with high spatial resolution over wide swaths and under all weather conditions. In combination with other information sources like Automatic Identification System (AIS) data fusion products are available to support the Maritime Situational Awareness.

The Sea Sate Processor (SSP) was developed for fully automatic processing of high-resolution Synthetic Aperture Radar (SAR) data from TerraSAR-X (TS-X) satellites and implemented into the processing chain for Near Real Time (NRT) services in the DLR Ground Station "Neustrelitz". The NRT chain was organised and tested to provide the processed data to the German Weather Service (DWD) in order to validate the new coastal forecast model CWAM (Coastal WAve Model) in the German Bight of the North Sea with 900 m horizontal resolution. The NRT test-runs, wherein the processed TS-X data were transferred to DWD and then incorporated into forecast products reach the best performance about 10 min for delivery of processed TS-X data to DWD server after scene acquisition. To do this, a new empirical algorithm XWAVE_C (C = coastal) for estimation of significant wave height from X-band satellite-borne SAR data has been designed for coastal applications. The algorithm is based on the spectral analysis of subscenes and the empirical model function yields an estimation of integrated sea state parameters directly from SAR image spectra without transformation into wave spectra. To provide the raster coverage analysis, the SSP intends three steps of recognising and removing the influence of non-sea-state-produced signals in the Wadden Sea areas such as ships, buoys, dry sandbars as well as nonlinear SAR image distortions produced by e.g. short and breaking waves. For the validation, more than 150 TS-X StripMap scene sequences with a coverage of ~30 km × 300 km across the German Bight since 2013 were analysed and compared with in situ Buoy measurements from 6 different locations. On this basis, the SSP autonomous processing of TS-X Stripmap images has been confirmed to have a high accuracy with an error RMSE = 25 cm for the total significant wave height.

The presentation focus on the framework and product development as an example for near real time applications for the maritime domain. Applications to derive maritime value added products like ship detection based on remote sensing SAR image data from TerraSAR-X are presently developed and integrated at the ground station Neustrelitz, part of the German Remote Sensing Data Center.

Since 2007, the TerraSAR-X mission provides high-resolution SAR data with high quality and an unprecedented geometric accuracy to both commercial and science users. The first satellite TSX was supplemented by a second, mostly identical, satellite TDX in 2010 for the generation of a global Digital Elevation Model (DEM) in the frame of the TanDEM-X (TerraSAR-X Add-On For Digital Elevation Measurements) mission. Since then the two missions share a joint space segment consisting of both TSX and TDX and a common ground segment which was first developed for TerraSAR-X and had been later extended for TanDEM-X. While TanDEM-X uses both satellites (nominally one in receive-only mode) for an acquisition, TerraSAR-X data are acquired by either one of the two satellites. It is this sharing of one ground segment which necessitated on-going updates of the TerraSAR-X ground segment in accordance with the TanDEM-X mission constraints. Furthermore, both satellites show a good health and resources, specifically the TSX which has considerably exceeded its nominal life time of 5.5 years by now, and thus justify on-going improvements to even better support user needs. This presentation summarizes the current SAR imaging modes and product portfolio and describes the recent ground segment system updates. Among these are the following: After having completed the global DEM data acquisition, the TanDEM-X mission currently focuses on the provision of radar data products for a number of new science technology related applications based not only on a bistatic, but also on a pursuit mono-static flight configuration. This pursuit mono-static flight configuration with its time lag of about 10 seconds between the two satellites allowed the processing of TanDEM-X acquisitions into TerraSAR-X product pairs which are offered to both the TerraSAR-X and the TanDEM-X science community. This includes fully polarimetric and along-track interferometry data acquired in the dual-receive antenna configuration. Receiving stations may be combined in a so-called TerraSAR-X ground station pool. Data takes ordered for such a station pool (instead of a single receiving station) are then planned for downlink to one or even several stations (in case of a partial downlink) by the mission planning leading to a much better exploitation of the limited downlink resources. Originally, the TerraSAR-X mission was not designed for near-real time (NRT) applications. Thus, the NRT functionality had to be included “as best as possible” within given constraints. Due to a growing demand for TerraSAR-X NRT products, the NRT capability was extended considerably over the last years. The TerraSAR-X product portfolio now supports maritime applications like ship and oil detection, wind and wave products are currently under integration. An improved mission planning and the downlink station pool concept leads to reduced product latencies.

TerraSAR-X, the first national German radar satellite, was launched in June 2007. It carries an X-band high-resolution synthetic aperture radar instrument featuring Stripmap, ScanSAR and, particularly, Spotlight imaging in a variety of different polarization modes. The mission completed its commissioning phase (CP) in December 2007, the provision of the SAR products for both the scientific and commercial user community was started in January 2008. One central TerraSAR-X element on ground is the pay-load ground segment PGS. From the beginning of the mission, PGS was nominally operated. About ten thousand data takes were already acquired and processed in 2007, not only for SAR verification and calibration purposes, but also for the operational ground segment validation. This paper provides the commissioning and early operational phase results from the SAR payload data processing perspective addressing data reception and SAR processing. Specifically the tuning and adjustment of the TerraSAR-X multi-mode SAR processor TMSP to meet the in-orbit data characteristics and to optimize the SAR focusing results is addressed. Relevant issues are the high-bandwidth chirp replica processing, side lobe suppression, Doppler frequency determination, processor normalization and phase-preservation.

TerraSAR-X, the first national German radar satellite, was launched in June 2007. It carries an X-band highresolution synthetic aperture radar instrument featuring the operational imaging modes Stripmap, ScanSAR and, particularly, Spotlight in a variety of different polarization modes. The TerraSAR-X mission completed its commissioning phase in December 2007 and started the provision of high-resolution products from advanced SAR modes for both the scientific and commercial user community from January 2008 on. The payload ground segment (PGS) is responsible for the reception of the SAR payload data, their archiving and processing and the distribution of the generated SAR basic products to users. From the first mission day on, PGS was operated successfully using both its request workflow starting with the input of user orders and ending with the delivery of the generated product and its operational SAR data workflow for the reception and processing of the SAR payload data. The central part of PGS is the TerraSAR Multi-Mode SAR Processor (TMSP) focusing the SAR data in a unified way for the different imaging configurations. A wide range of processing options spanning from phase preserving complex products in slant range geometry to orthorectified terrain corrected intensity images lead to a comprehensive collection of SAR product types and variants. Together with instrument calibration, checkout and tuning of the SAR processor was one of the major tasks of the commissioning phase with respect to product verification and validation. This paper summarizes the key activities and exposes the achievements obtained with respect to focusing quality.

The successful launch of the second national German radar satellite TDX-1 in June 2010 marked the beginning of the challenging TerraSAR-X add-on for Digital Elevation Measurement mission. Major TanDEM-X mission goal is the consistent generation of a worldwide global digital elevation model with high accuracy. Therefore the two satellites TSX-1 and TDX-1 are flown in close configuration to form a single-pass space-borne SAR interferometer. To counterbalance the interferometric usage of the TSX-1 satellite, data for the ongoing TerraSAR-X mission have to be acquired by both satellites TSX-1 and TDX-1. Commissioning of the joint TerraSAR-X/TanDEM-X system was done in well-defined stages. TDX-1 was launched at June 21, 2010, and started its drift towards TSX-1. At July 22, 2010, it had reached a distance of 20 km and started its helix formation flight. This pursuit monostatic configuration was kept for 77 days to allow not only a thorough TDX-1 system verification, but also to qualify the overall system e.g. for the close formation flight with its zero-mean along track distance which started mid of October 2010. The good performance of both the TDX-1 satellite and the joint TerraSAR-X/TanDEM-X ground segment allowed to acquire a significant amount of TerraSAR-X data from the TDX-1 satellite in addition to the planned commissioning data set. These data represent nominal TerraSAR-X acquisitions and their processing thus leads to TerraSAR-X basic products showing a nominal product performance. This lead to the decision, that – after a careful inspection – these TDX-1 data will be made available to users for catalogue ordering despite their commissioning phase character. Of specific user interest among them are those scenes which were acquired with both the TSX-1 satellite and the TDX-1 satellite following each other in less than 3 seconds distance. This paper gives an overview of the different types of TerraSAR-X data acquired by both TSX-1 and TDX-1 in this unique pursuit monostatic flight configuration. It furthermore provides exemplary evaluations, e.g. in the area of ship detection.