The dAIEDGE Network of Excellence (NoE) seeks to strengthen and support the development of a dynamic European cutting-edge Artificial intelligence (AI) ecosystem under the umbrella of the European Lighthouse for AI, and to sustain the development of advanced AI. dAIEDGE fosters the exchange of ideas, concepts, and trends on cutting-edge next generation AI, creating links between ecosystem actors to help both the European Commission (EC) and the European Union (EU) and the peripheral AI constituency identify strategies for future developments in Europe. Our main objective is to advance Europe’s innovation and technology base by developing a comprehensive policy and governance approach to AI in order for the EU to become a world leader in innovation in the data economy and its applications.

We present a new approach for sea ice mapping based on Synthetic Aperture Radar (SAR) data from Sentinel-1 and an existing sea ice classification using optical-thermal data based on Sentinel-3. SAR and optical-thermal sensors provide different information about the sea ice situation: while SAR backscatter depends mainly on the topography of the sea ice surface and properties of the ice volume, optical sensors provide further information about the structure and moisture of ice and snow. In order profit from both sensors, a convolutional neural network (CNN) is trained with collocated images from both satellite missions. Compared to a pure SAR classification, the results of the combined approach show an improved classification reliability, especially in areas with open water.

This paper introduces an improved system for sea state observations for Near Real-Time (NRT) services using satellite-borne synthetic aperture radar (SAR). The empirical algorithm SAR-SeaStaR (SAR Sea State Retrieval) applies a combination of a classical approach using linear regression (LR) with machine learning (ML). SAR-SeaStaR includes a series of filtering and control procedures and a series of LR and ML model functions for different satellites/modes and following integrated sea state parameters: total significant wave heightHs, wave heights of dominant and secondary swells and windsea, mean, first and second moment wave periodsTm2, and windsea period. SAR scenes are processed in raster format, the output are fields for each parameter showing their spatial distribution. In the scope of this study, the ML models were developed forHsandTm2and implemented into SAR-SeaStaR for processing Level-1 products of X-band TerraSAR-X (TS-X) StripMap (SM) and C-band Sentinel-1 (S1) Interferometric Wide Swath Mode (IW), S1 Extra Wide (EW). The validations are based on processing large worldwide archives with several years of acquisitions. Hindcast data from numerical spectral models andin-situbuoys measurements are used as ground truth. The root mean squared errors of the complete system reached from these archived data for Hs are RMSE=0.35 m for TS-X SM (pixel spacing ca. 1.2-4.5 m pixel), RMSE=0.25 m for S1 Wave Mode (WV, ca. 3.5 m pixels), RMSE=0.42 m for the coarser S1 IW (10 m pixels) and RMSE=0.52 m for S1 EW (40 m pixels). SAR-SeaStaR was implemented in the Sea State Processor software using modular architecture and applied at the DLR Ground Station in Neustrelitz as part of an NRT demonstrator service. S1 IW data acquired over North and Baltic Sea are processed automatically, surface wind and sea state parameters are provided daily.

PERTEO is a mission proposal intended to provide real-time EO services to reduce Natural Disaster impact by proposing a heterogeneous small satellite constellation and performing in-orbit processing. For that aim, this mission takes advantage of the enhanced capabilities of AI edge processing by enabling on-demand user services through the "satellite as a service" concept. The design of the constellation includes three types of sensors: SAR, Hyperspectral and Multispectral imager VHR combining their capabilities to achieve almost real-time due to their in-orbit distances. Six spacecrafts are considered in each orbit organized in pairs (180° phased platforms) mounting the same instrument. The proposed solution is responsive achieving almost real-time latencies ( Ex: ≲ 1 min, down to seconds is achieved in the first observation product). Persistence is achieved with a low revisit time (≲ 1 hour any payload; ≲ 3 hour s any specific payload choice).

Satellite remote sensing acquisitions are usually processed after downlink to a ground station. The satellite travel time to the ground station adds to the total latency, increasing the time until a user can obtain the processing results. Performing the processing and information extraction onboard of the satellite can significantly reduce this time. In this study, synthetic aperture radar (SAR) image formation as well as ship detection and extreme weather detection were implemented in a multiprocessor system on a chip (MPSoC). Processing steps with high computational complexity were ported to run on the programmable logic (PL), achieving significant speed-up by implementing a high degree of parallelization and pipelining as well as efficient memory accesses. Steps with lower complexity run on the processing system (PS), allowing for higher flexibility and reducing the need for resources in the PL. The achieved processing times for an area covering 375 km2were approximately 4 s for image formation, 16 s for ship detection, and 31 s for extreme weather detection. These developments combined with new downlink concepts for low-rate information data streams show that the provision of satellite remote sensing results to end users in less than 5 min after acquisition is possible using an adequately equipped satellite.

In modern society, the ever-increasing demand for Earth Observation products in a large variety of sectors is exposing the limitations of traditional satellite data chain architectures. The European Union Horizon 2020 EO-ALERT project aims at overcoming the existing bottlenecks by leveraging the performance of state-of-the-art commercial off-the-shelf devices to move the critical elements of data processing on the flight segment without sacrificing processing performance. This paper introduces the architecture of the EO-ALERT CPU Scheduling, Compression, Encryption and Data Handling Subsystem, responsible for coordinating the onboard optical and Synthetic Aperture Radar data chains, as well as providing data compression, encryption, and storage services. The performance obtained by a reference implementation of the proposed architecture is also presented, showing an extremely low contribution to the overall system latency that allows real-time Earth Observation product delivery to the end user in less than 5 min.

This paper presents the architecture and results achieved by the EO-ALERT H2020 project. EO-ALERT proposes the definition and development of the next-generation Earth Observation (EO) data processing chain, based on a novel flight segment architecture that moves optimised key EO data processing elements from the ground segment to onboard the satellite, with the aim of delivering the EO products to the end user with very low latency (in almost real-time). This paper presents the EO-ALERT architecture, its performance and hardware. Performances are presented for two reference user scenarios; ship detection and extreme weather nowcasting/monitoring. The hardware testing results show that, when implemented using Commercial Off-The-Shelf (COTS) components and available communication links, the proposed architecture can deliver EO products and alerts to the end user with a latency lower than one-point-five minutes, for both SAR and Optical Very High Resolution (VHR) missions, demonstrating the viability of the EO-ALERT concept and architecture.

This paper describes a prototype implementation of ship, wind, and sea state detection algorithms for satellite on-board SAR processing designed for Maritime Situation Awareness. Existing algorithms were adapted to run on a Multi-Processor-System-On-Chip (MPSoC) combining an FPGA and an ARM CPU and further optimized for fast runtime on the system. The achieved processing times were 20 s for ship detection and 16 s for sea state detection on a 29 Mpx SAR image. SAR processing is one component of a larger prototype system being developed in the frame of the H2020 project EO-ALERT, which further comprises an optical data chain, data compression/encryption, and delivery on multiple MPSoC boards.

This study presents an extension to recent ship wake detectability models based on SAR image analysis with machine learning. In contrast to our previous works, we model the detectability of certain wake components individually. The underlying data set is obtained by extracting possible ship wake signatures from SAR imagery by collocation with Automatic Identification System data. The developed detectability models are based on machine learning. They generally confirm previous findings based on simulated SAR data or qualitative image analysis. The results from our previous wake detectability model are compared to initial results from our new wake component detectability model.

The European Copernicus programme generates massive amounts of Earth Observation (EO) data, with the goal of improving our environmental understanding and management on local, regional, and global level. An initiative of this magnitude comes with a set of challenges, especially pertaining to effective service management once the huge volumes of raster data files are distributed across open data centers and commercial companies.The BigDataCube project responds to this challenge through the concept of federated, analysis-ready datacubes, exposed via the open OGC geo standard interfaces for interoperable access and processing, WMS, WCS, and WCPS.Such federated datacube services have been established in the project between the public German Copernicus hub, CODE-DE, and a commercial cloud provider, cloudeo AG, altogether offering access to more than 500 TB. Both data pools are federated in a location-transparent manner, establishing a common information space where users can query and combine datacubes without knowing their location of storage. In this nucleus of a growing federation recently the Alfred Wegener maritime research institute has joined, and further data centers are in the line. The spectrum of functionality available is showcased through several realistic use cases, including a value-adding Sentinel-1 SAR product on sea state and wind speed in the North Sea. The platform used is the pioneer datacube engine and OGC reference implementation, rasdaman. We report on the outcomes of the project, lessons learned, and further work foreseen in this area.

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.