Science Service System

Summary of Proposal LAN0229

TitleCrises management using high-resolution SAR-data
Investigator Ulander, Lars - Swedish Defence Research Agency, Radar Systems
Team Members
Rasmusson, Johan - Swedish Defence Research Agency, Radar Systems
Lundberg, Mikael - Swedish Defence Research Agency, Radar Systems
LicEE Frölind, Per-Olov - Swedish Defence Research Agency, Radar Systems
MSc Larsson, Björn - Swedish Defence Research Agency, Radar Systems
MSc Gustavsson, Anders - Swedish Defence Research Agency, Radar Systems
SummaryThe objective of the project is to evaluate and demonstrate the use of TerraSAR-X data for crises management. The application focus is on two scenarios with different requirements. Scenario 1 is evacuation of a city in a crises situation. Scenario 2 is control of a refugee camp and its forest surroundings. Both scenarios will be evaluated with one test site located in Sweden and one in Africa. The test sites in Sweden have been chosen close to our agency location in Linköping and where a controlled infrastructure and extensive data bases already exist, including high-resolution DEM (laser) and airborne sensor data. The first approach will be to control and document activities on the ground in the Swedish test sites and evaluate the detection performance using multi-temporal TerraSAR-X data as a function of operating conditions (incidence angle, aspect, polarization etc). The outcome of this work is curves which describe detection probability vs false-alarms. The next step is object characterization. It will also be possible to perform modelling and simulation of the radar backscattering based on the extensive data sets in these areas and compare with the SAR imagery. Algorithms for change detection and classification tools will be developed and evaluated. Multiple polarization and interferometry data will be evaluated. A total of about 50 images will be asked for. The second approach will be to demonstrate the capabilities of TerraSAR-X in two areas in Africa where only limited information is available. We plan to collect multi-temporal data over the same area during the major seasons but also within a season for evaluating scene stability. There may also be possibilities to perform a field trip to collect in situ data for evaluation purposes. TerraSAR-X data will be combined with other data sources and evaluated, i.e. high-resolution optical data (e.g. IKONOS). A total of about 50 images will be asked for. A related proposal has been submitted in parallel for Radarsat-2 data.
Detailed reportThe final report is available as a pdf and can be obtained upon request. I was unfortunately not able to upload the file. Below follows a summary. Introduction This report summarizes the work performed in the study “Efficient monitoring of antagonistic crises and force majeure events”. The study was a co-operation between the Swedish Defence Research Agency (FOI) and the Swedish Armed Forces (FM) during 2007 and 2008, where FM has been the end-user. The study has been partly funded by the Swedish National Space Board, SNSB (Rymdstyrelsen). A main objective of the study has been to evaluate and demonstrate the usefulness of very high resolution SAR images. Within the next few years, there will be several civilian radar satellite systems available with everyday, all weather capability and a resolution near 1 meter. They will be more useful for a broader range of remote sensing applications because of their much higher resolution (near 1 m) compared to previous radar satellite generations (10-100 m resolution). Even higher resolution, down to dm, already has been demonstrated in airborne SAR images. This report has focused on synthetic aperture radar (SAR) images from the German aerospace centre’s (DLR) radar satellite TerraSAR-X, via an approved proposal to DLR from FOI, “Crises management using high-resolution SAR-data”. Using time series of SAR satellite images and automatic change detection (CD) of changes in e.g. infrastructure (for instance buildings and roads), and location of specific features (such as vehicles and aircraft) have been investigated with promising results. Using SAR satellite images and CD, long time civilian development and military escalation can be monitored, as well as more rapid changes in infrastructure, buildings and land topography, caused by environmental changes, natural disasters and other force majeure events. Maps may rapidly be generated with updated information for areas of interest. The ability to establish a good situational awareness is strongly dependent on the allowed preparation time as well as the availability of sensor systems and other resources for information and intelligence collection on site or via remote sensing. Situational awareness information may, simplified, be divided in two parts; I) Static or slowly varying background data such as topography and functionality of building and places. II) Dynamic data, varying over time, for instance positions of vehicles. User needs The Swedish Armed Forces (FM) often needs remote sensing and other methods to gather information for effective mapping and planning of missions. Typical questions for FM may be: • How is the terrain situated? • What kind of infrastructure is available, e.g. airports, harbor, railway, roads, bridges and power grids? • May the infrastructure safely be used for troops and heavy transports? • Where may threats be expected? Which routes should be used or avoided? • Where may a camp safely be located? • How should new maps efficiently be produced in areas without or with out of date existing map information? • How should changes and damages effectively be monitored and documented over time in the area of interest? • How to make missions as safe and efficient as possible? Studied areas in Africa and Sweden Together with FM, we have identified some areas of interest, including an international hot spot; Abéché, Tchad with urban environments and refugee camps located in a semi-arid landscape near the boarder of the Darfur province in Sudan. We have also included four Swedish areas: Stockholm Norrrköping, Linköping and Kvarn (army training facility located 40 km N.W. of Linköping), where the ground conditions are controlled and well documented. Extensive data bases already exist for Kvarn and Norrköping, including high-resolution laser digital elevation maps (DEMs) and airborne sensor image data from laser, VIS, IR and SAR. For the areas mentioned above, we have obtained high resolution SAR images from DLR’s radar satellite TerraSAR-X via the approved proposal“Crises management using high-resolution SAR-data”. Summary and conclusions In summary, high resolution SAR satellite images rapidly provide valuable information for large areas and can used to effectively pinpoint smaller areas of interest from larger data sets. The SAR image information relates to how the terrain is situated, the presence and status of available infrastructure and buildings. SAR images can be used to monitor changes and damages over time in areas of interest in most part of world. The SAR technique is independent of daylight and bad weather conditions. SAR images are very suitable for automatic image processing methods, e.g. the use of change detection to extract a few interesting spots from high resolution images covering large areas. Further studies may then be focused on these interesting spots which likely will save resources and time. SAR images and change detection methods may be combined with other remote sensing methods such as high resolution optical images for efficient mapping, planning and monitoring of crises and missions. The following conclusions are drawn from our studies: • In terms of usefulness for management of force majeure events and antagonistic crises, the arrival of the next generation high resolution SAR civilian satellites with near 1 m resolution, which is about 10 times higher than previous generations, is a major step forward enabling many new exciting applications. • Optical images may still when available be the best alternative for a human image analyst to recognize and identify a specific object, provided that he or she has previous knowledge in which limited area to look. • When time, human or other resources are limited, there is growing need to use automatic computerized method to pinpoint smaller interesting areas. • SAR image change detection is an automatic method that can be used to suppress stationary objects and efficiently detect significant changes in small interesting areas from large images, e.g. movement of vehicles. • Pixel-based change detection, based on processing of pairs of high resolution SARimages, is a technique that can be used to find interesting areas for further manual or automatic analysis. • SAR images and change detection methods should be combined with other remote sensing methods for efficient mapping, planning and monitoring of crises and missions.

Back to list of proposals

© DLR 2004-2016