Following the huge success of the 2019 event in Rome, the SAR Analytics Symposium 2022 once again featured analytical thought leaders and provided a unique meeting place for SAR service providers and applied SAR analytics consumers from across various industries.

Attendees gleaned knowledge and ideas from other successful real-world analytical applications, all while sitting in a truly beautiful setting in the middle of Berlin, renowned for its exceptional variety of attractions, its flourishing cultural scene and a way of life that's both fast-paced and relaxed.

Presentations from Thursday, October 20

Click Here for Presentations from Friday, October 21


SESSION 1


Mack Koepke
Capella Space

VHR SAR imagery - timely and reliable - for a fast-changing world
Abstract

VHR SAR imagery - timely and reliable - for a fast-changing world

Mack Koepke - Capella Space


The requirements for remote sensing data providers in terms of spatial and temporal resolution are growing with the technical capabilities of modern small satellite constellations. Capella will provide a live demonstration of how easily, quickly, and effectively Capella VHR SAR data can be tasked and downloaded directly by the customer via their Image-Tasking platform Console. This service enables near-real-time tasking of the constellation with no human in the loop, translating into full automation of task scheduling, secure and private customer operations, and speedy data processing and delivery to enable access to global monitoring where and when you need it. We will in detail explain what we mean with: - High-Quality - Very high resolution and low noise leads to enhanced image clarity - Timely - Rapid fully-automated order-to-delivery means faster speed to insight - Frequent - Increasing high-cadence revisit timeframes as our constellation grows - Accessible - Intuitive online platform with self-serve catalog search, ordering and tasking Capella’s high-resolution SAR satellites are matched with unparalleled infrastructure to deliver reliable global insights that sharpen our understanding of the changing world – improving decisions about commerce, conservation, and security on Earth. Learn more at www.capellaspace.com.

Presentation PDF

Nicolas Longepe
ESA

Advanced Solutions for SAR processing and analytics using AI
Abstract

Advanced Solutions for SAR processing and analytics using AI

Nicolas Longepe - ESA


The challenge of scaling commercial remote sensing systems is bridging the gap between existing markets and new markets that not only need new solutions but process information in a unique way. Capella Space is bringing reliable and timely SAR data sets to commercial markets with unprecedented accessibility and quality. To meet new markets and customers where they are today, we will have to find ways to deliver the information embedded in our data sets in a way that is domain appropriate and operationally feasible. This presentation will be about finding the value in SAR and delivering it to the market sectors who are not yet committed users of remote sensed data, let alone SAR-enabled information streams.

Presentation PDF

Dr. Paolo Pasquali
sarmap SA

SAR Image Simulation and Automatic Target Recognition
Abstract

SAR Image Simulation and Automatic Target Recognition

Dr. Paolo Pasquali - sarmap SA


SAR images are not necessarily very easy to visually interpret, and these difficulties can even increase in case of images of man-made objects and infrastructures. On the other hand, the exploitability of automated methods for spatial features’ identification and recognition deeply relies on the availability of enough reference images to train recognition systems, when based on Artificial Intelligence and Deep Learning methods. This presentation will introduce the results of the development of a SAR image simulator and of an Automatic Target Recognition system tightly linked to it. The focus of the first component is in generating simulated SAR images that provide a rigorous representation of how a 3D scene would look like if acquired from a given SAR sensor: the generated images can be exploited for the training of human operators and as support to visual interpretation, as well as for the training of the ATR component. This second system is providing accurate automated detection and classification of objects and spatial features in a SAR image.

Presentation PDF

Dr. Oliver Lang
Airbus

Automatic Object Recognition in High Resolution SAR imagery
Abstract

Automatic Object Recognition in High Resolution SAR imagery

Dr. Oliver Lang - Airbus


The presentation introduces an operational AI/ML solution for very fast object detection and recognition in high resolution SAR imagery. Focus will be on the training aspects and on the discussion of capabilities and challenges.

Presentation PDF


SESSION 2


 

Skevi Perdikou
Geofem limited

Addressing today’s infrastructure challenges using SAR data
Abstract

Addressing today’s infrastructure challenges using SAR data

Skevi Perdikou - Geofem limited


Climate change is causing an increased occurrence of geohazards (e.g., landslides, sinkholes) with enormous economic impact and disruption, affecting infrastructure globally. We present the use of SAR combined with geotechnical data and AI for geohazard susceptibility assessment. This allows problem areas to be pinpointed early and trends predicted. This, in turn, leads to earlier maintenance interventions which are more economical, less disruptive and build greater resilience to climate change back into our infrastructure. As well as infrastructure, this is also being applied for the benefit of the construction, insurance, real estate, energy and mining sectors.

Presentation PDF

Ralph Schmidt
State Bureau of Surveying and Geoinformation SH

SAR4Infra - traffic infrastructure monitoring using Sentinel-1 InSAR
Abstract

SAR4Infra - traffic infrastructure monitoring using Sentinel-1 InSAR

Ralph Schmidt - State Bureau of Surveying and Geoinformation SH


Traffic infrastructure plays a vital role in today's society. For security and economic reasons, it is essential to monitor traffic infrastructure, including roads, railways, and bridges. Assessing the health of traffic infrastructure with conventional geodetic approaches like levelling or GNSS observations is time-consuming, costly, and limited to sparse locations. In contrast, Earth observation data from satellite missions provide unique opportunities for spatially wide-scale and temporally high-resolution monitoring. In the scope of our project SAR4Infra we develop an automatic InSAR processing chain to generate a risk displacement map for traffic infrastructure in Schleswig-Holstein. We investigate the combination of different time series techniques to obtain a dense distribution of pixels on the traffic infrastructure. The reliable pixels consist of persistent and distributed scatterers. We improve the signal-to-noise ratio of distributed scatterers by phase-linking and subsequently process both persistent and distributed scatterers jointly in a persistent scatterers time series approach. To decrease the computational burden, we estimate the deformation parameters sequentially once a new SAR image is acquired without re-estimating the whole time series. We will present the current status and prospects of the SAR4Infra project and highlight the challenges and opportunities of Sentinel-1 InSAR for traffic infrastructure monitoring.

Presentation PDF

Christin Lubitz
Airbus

Bridge Deformation Monitoring: Fusion of InSAR, Geodesy and Simulations
Abstract

Bridge Deformation Monitoring: Fusion of InSAR, Geodesy and Simulations

Christin Lubitz - Airbus


Satellite-based InSAR deformation measurements offer new capabilities in the context of Structural Health Monitoring (SHM). The presentation discusses a monitoring concept combining InSAR, geodetic measurements and structural simulations.

Presentation PDF

Presentations from Friday, October 21


SESSION 3


Jeff Hurley
MDA

Maritime Insights: Space-based data and analytics
Abstract

Maritime Insights: Space-based data and analytics

Jeff Hurley - MDA


Illegal, Unreported and Unregulated (IUU) Fishing is one the most pressing issues facing our natural world and is estimated to cost the global economy in excess of $23 billion annually or about 20%-30% of the global fish catch depending on the market. This challenge is particularly difficult to address due to the remote areas where IUU Fishing is occurring. MDA is a leader in the use of Synthetic Aperture Radar to extract analytics about vessels on the high seas far from shore to tackle the challenge of IUU Fishing. MDA will present the latest research and development work it has operationalized for extraction of vessels from SAR, vessel velocity and vessel classifications. MDA will review the analytical and machine learning approaches employed to enable these new capabilities. Further, this presentation will provide insight on how MDA investments in SAR fusion with other sources such as RF, VIIRS and optical is enabling fisheries intelligence users to address their most difficult maritime domain awareness challenges such as tracking illicit dark vessel activity.

Presentation PDF

Patrick O'Keeffe
NATO Centre of Excellence for Operations in Confined and Shallow Waters (COE CSW)

SAR for Maritime Situational Awareness (MSA)
Abstract

SAR for Maritime Situational Awareness (MSA)

Patrick O'Keeffe - NATO Centre of Excellence for Operations in Confined and Shallow Waters (COE CSW)


SAR for Maritime Situational Awareness (MSA) with examples of use cases in multinational trials and exercises with NATO and non-NATO partners. SAR is one element of the digitalisation of the oceans.

Presentation PDF

Jaewoo Shin
Open Nuclear Network

Leveraging SAR for Nuclear Fuel Cycle Monitoring
Abstract

Leveraging SAR for Nuclear Fuel Cycle Monitoring

Jaewoo Shin - Open Nuclear Network


SAR provides unique opportunities and challenges for monitoring nuclear fuel cycles for nuclear safeguards purposes and research on various weapons of mass destruction (WMD) programmes around the world. This presentation illustrates some of these opportunities and challenges with case studies from the DPRK context.

Presentation PDF

Giulia Tessari
sarmap SA

Machine Learning techniques using multi-sensor data for groundwater potential estimation
Abstract

Machine Learning techniques using multi-sensor data for groundwater potential estimation

Giulia Tessari - sarmap SA


Chadian population is suffering of the effect of climate change due to a severe water scarcity in African arid regions. Groundwater potential maps might contribute to the hydrogeological knowledge and support the water resources management to plan the exploitation of the aquifers in Chad. A groundwater potential map based on a machine learning approach is presented for the crystalline aquifer of the eastern Chad. The spatial-distributed explanatory variables are derived from multi-sensor data: multitemporal Landsat-8 and Sentinel-2 optical data, Sentinel 1 SAR data, Meteosat Second Generation (MSG) TAMSAT and GLO-30 Digital Elevation Model are used to generate seasonal and static products. Geological and hydrogeological fieldwork and knowledge-based information completed the analysis. A sample of 488 wells and boreholes are used to train and test the machine learning classifiers to define the presence of groundwater. Analysis on the performance of the supervised classification algorithms shows that Random Forest provides the higher test score, the analysis on the most relevant explanatory variables shows that fracture density, slope, SAR coherence (interferometric correlation), topographic wetness index, basement depth, and distance to ephemeral channels have the higher probability to predict the groundwater occurrence. The results allowed to identify the wadis and piedmont region in Chad as the most favourable area for groundwater exploitation.

Presentation PDF

 

Gianni Onorato
CNR IREA

Multi-temporal/multi-frequency satellite SAR data P-SBAS analysis of natural & builtup environments
Abstract

Multi-temporal/multi-frequency satellite SAR data P-SBAS analysis of natural &smp; builtup environments

Gianni Onorato - CNR IREA


In the last decades the need to effectively investigate deformation phenomena affecting wide areas of the Earth surface has fostered the exploitation of satellite SAR images and advanced Differential SAR Interferometry (DInSAR) techniques. In this framework, the Parallel Small BAseline Subset (P-SBAS) DInSAR approach has demonstrated, by exploiting HPC infrastructures also including GPU-based architectures, its capability to efficiently perform massive surface displacement analyses related to natural and anthropic hazard scenarios, by generating spatially dense displacement time series with millimeter accuracy and at different spatial resolutions. In this work, we analyze the effectiveness of the P-SBAS DInSAR approach to study, at different spatial scales, deformation phenomena affecting the natural and the built-up environment. To do this we first show the P-SBAS capability to investigate the built-up environment of some of the largest Italian cities by processing full resolution X-band stripmap SAR data stacks acquired by the Italian first and second generation COSMO-SkyMed constellation. Subsequently, we show some results achieved by exploiting large volumes of the C-band SAR data acquired through the TOPS mode of the European Sentinel-1 sensors, relevant to the natural environment of the Italian territory. The final analysis is focused on the results obtained on selected stacks, processed through the P-SBAS approach, of L-band stripmap SAR data acquired by the SAOCOM-1 Argentinian constellation.

Presentation PDF

Dr. Salvatore Stramondo
INGV

Protecting infrastructures from natural and anthropogenic hazards
Abstract

Protecting infrastructures from natural and anthropogenic hazards

Dr. Salvatore Stramondo - INGV


Thanks to the wide availability of SAR satellite missions, Multitemporal Synthetic Aperture Radar Interferometry (MT-InSAR) are nowadays reliable techniques to measure the effects of natural phenomena (earthquakes, landslides, etc.) and anthropogenic activities (exploitation of groundwater, hydrocarbons, gas storage, etc.) on urban areas and infrastructures (railways, airports, bridges, etc.). INGV GEOSAR Lab. exploited several commercial and open-source softwares such as SARScape, GAMMA and StaMPS to generate MT-InSARtime series stemming from C-Band Sentinel-1 (S1) and X-band COSMO-SkyMed (CSK) datasets. Finally the European Union released the Programme for Critical Infrastructure Protection (EPCIP) defining the overall framework for activities aimed at improving the protection of critical infrastructure in Europe. EU started the European Infrastructure Simulation and Analysis Centre (EISAC) initiative, for supporting Operators and Public Authorities in better protecting assets and in enhancing their resilience with respect to all hazards. In Italy, EISAC.it results from a joint ENEA and INGV collaboration agreement established in 2018.

Presentation PDF

Robert Siegmund
GAF AG

Large scale ground deformation services and their potential for downstream applications
Abstract

Large scale ground deformation services and their potential for downstream applications

Robert Siegmund - GAF AG


Picking up recent trends in high volume and large scale interferometric data analysis, basic ground deformation services were implemented on a national (Germany) and Pan European scale. As a service provider, GAF utilises a robust and validated processing approach deployed on its private cloud infrastructure. Concepts of our basic service operation plus interfaces for value adding and downstream applications will be presented.

Presentation PDF

 

 


SESSION 4


Miguel Castro
IABG

Comparison of SAR sensors in emergency applications in the context of La Palma eruption (2021)
Abstract

Comparison of SAR sensors in emergency applications in the context of La Palma eruption (2021)

Miguel Castro - IABG


The volcanic eruption on the island of La Palma (Canary Islands) started on 19.09.2021, specifically as a flank eruption at the Cumbre Vieja volcanic ridge comprising the southern half of the island. On 13.12.2021, volcanic activity stopped and was finally declared to have ended by 25.12.2021. With a duration of 85 days, it has been the longest known eruption on La Palma and the first on the island since the eruption of Teneguía in 1971. The eruption caused the evacuation of around 7,000 people and the lava flow covered an area of about 3.5 km in its widest point and 6.2 km long, affecting over 1,200 hectares. It reached the sea in two different locations forming two new peninsulas. During the eruption, approximately 3,000 buildings and 70 km of transportation infrastructure were destroyed. Towns such as Todoque or Los Campitos were significantly affected, as well as agricultural land, destroying 370ha of planted areas and affecting immensely the economic situation of local farmers.

In this talk, an overview of the emergency and main outcomes derived from the processing of SAR data will be presented. The results were produced by IABG in the context of the Copernicus Emergency Management Service in support of the Spanish authorities.

Presentation PDF

Filippo Britti
e-GEOS

When polarimetry meets very high frequencies: COSMO Second Generation
Abstract

When polarimetry meets very high frequencies: COSMO Second Generation

Filippo Britti - e-GEOS


The COSMO-SkyMed Second Generation [CSG] is at the forefront of radar technologies and ensures improvements and guarantee continuity with the first generation satellites, preserving the high quality and the highest precision features, both required for the interferometric activities. The programme is funded by Agenzia Spaziale Italiana (ASI), the Italian Ministry of Defence and the Italian Ministry of Education, Universities and Scientific Research. Thales Alenia Space is the prime contractor responsible for the construction of two satellites, while Telespazio provides the ground segment and offers integrated logistics for operations. The present work focusses on the enhanced performances of second generation radar data and their new application potentials. In details, it will be showed how the polarimetric capabilities of the new generation radar sensors, added to the high frequency revisit of the constellation, enhance Earth Observation performances of the COSMO mission. e-GEOS, world-wide exclusive distributor of COSMO-SkyMed and operator, on behalf of ASI, of the IC-UGS (Italian Civilian User Ground Segment) at the Matera Space Center, is one of the most important satellite data providers on the market, providing also several Earth Observation based services. Mainly focusing on providing and using Synthetic Aperture Radar (SAR) images, e-GEOS is one of the major actors in the field of Emergency Services, Interferometric Analysis, Defense and Intelligence analysis, and many other services that have been developed during the years and that take advantage of the new missions available on the market.

Daniel Carrasco
HISDESAT

Benefits of Very High Resolution X-band data for INSAR purposes
Abstract

Benefits of Very High Resolution X-band data for INSAR purposes

Daniel Carrasco - HISDESAT


ONDA is the EU funded DIAS (Data and Information Access Service) led by Serco Italia that brings together data and information close to processing facilities in a cloud-based environment. The platform provides a new way to process massive amount of data from different sources and to apply analytics that lead to the development of tailored EO value added products.

The whole archive of Sentinel-1 products as well as ENVISAT ASAR are available. Complemented by optical datasets, ONDA gives free access to multi petabyte of data worldwide. VHR SAR and optical data are also commercially available to offer a larger range of spatial and spectral resolution. Today the platform has more than 1000 registered users and is close to support more than 100 applications, from Research and Development to commercial services. CNR IREA is one of these users.

The application presented by CNR IREA studies the temporal evolution of ground deformation over large portion of the Earth surface. DInSAR analysis is performed at continental scale to monitor surface deformation and estimate the corresponding deformation velocity. 72000 S1 products (IWS) for an area of more than 4.5 million sqkm are analysed, demonstrating the suitability of ONDA for big data application.

Presentation PDF

Naruo Kanemoto
Space Shift, Inc.

Improving SAR satellite capability with AI/ML, information extraction from SAR “Data" not only from amplitude image
Abstract

Improving SAR satellite capability with AI/ML, information extraction from SAR “Data" not only from amplitude image

Naruo Kanemoto - Space Shift, Inc.


Space shift is a Japanese company based in Tokyo develops SAR satellite data analytic software with AI/ML. Analysis of SAR data should not be limited to its amplitude images as it contains more information such as phase changes in lower products like SLC. We developed our own AI/ML method to utilize hidden features to improve its prediction accuracy to fit market requirements. In this presentation we will introduce our methodology to utilize all of available “Data” contained inside of SAR satellite signals to improve its capability for many type of civil applications and actual cases of our various customers.

Presentation PDF

Thaïs Bendixen
UP42

Operationalise SAR data collection and processing at scale
Abstract

Operationalise SAR data collection and processing at scale

Thaïs Bendixen - UP42


SAR data is fragmented and spread across multiple provider platforms, which makes It hard to build and distribute algorithms, especially when you want to do this at scale. UP42 provides access to several SAR data providers, including Capella Space, TerraSAR-X and Sentinel-1. We also provide a workflow engine and scalable infrastructure, which means we enable you to bring your own algorithms to the platform, to process SAR data in our platform. You can build processing pipelines and use already existing SAR-processing algorithms, such as ship detection and ground displacement. In this presentation we want to demonstrate some examples of SAR data processing at scale.

Presentation PDF