Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal LAN0712

TitleInterferometric Synthetic Aperture Radar (SAR) for investigating the terrain morphology, land cover and elevation of three sub-Antarctic Islands
Investigator Mitchell, Anthea - University of New South Wales, School of Biological, Earth and Environmental Sciences
Team Member
A/Prof Ge, Linlin - University of New South Wales, School of Surveying and Spatial Information Systems, Faculty of Engineering
Dr Laffan, Shawn - University of New South Wales, School of Biological, Earth & Environmental Sciences
SummarySummary and objectives: This project will investigate methods of DEM generation, coastline detection and terrain classification using Interferometric SAR (InSAR) data acquired over Australian Antarctic Territory. Accurate elevation data and land cover information is essential for topographic mapping, surface characterisation and monitoring in response to changing environmental dynamics and climate change. Satellite SAR provides the only cost-effective and efficient means of acquiring topographic information over largely inaccessible sub-Antarctic Islands. The derived DEMs and spatial map data will be made available to the Australian Antarctic Division (AAD) and the German Space Agency (DLR) for inclusion in their scientific databases and use in ongoing research projects. Methodology: Archive, multi-frequency InSAR data will be acquired over Macquarie, Heard and McDonald Islands at X-band (TerraSAR-X), C-band (Radarsat) and L-band (ALOS PALSAR and JERS-1). SRTM data from 2000 will be downloaded from the NASA website. Conventional InSAR processing will be applied to generate DEMs for all three islands using the available SAR data. The best DEM for Macquarie Island will be compared to a DEM previously generated using AIRSAR TOPSAR data. The best DEM for Heard Island will be compared to an earlier DEM generated from Radarsat-1 data. DEMs will be differenced to evaluate any topographic change between dates. The potential of Differential InSAR (DInSAR) for identifying areas of possible land deformation will also be tested. A pilot study conducted by the investigators in 2008 demonstrated the use of InSAR for DEM generation and terrain classification using ALOS PALSAR data acquired over Macquarie and Heard Islands. Preliminary results indicated that the PALSAR data were suitable for DEM generation, and change detection with archive DEMs was possible. One aim of this project is to extend the InSAR DEM generation using TerraSAR-X data. The pilot study also determined that single polarisation L-band data was limited for terrain mapping. The project will investigate the integration of dual polarisation and multi-frequency (X- and L-band) data for improved discrimination of cover types. All SAR data will be processed to backscatter coefficient using standard radar equations. The coastline of each island will be detected using a combination of filtering, segmentation and thresholding techniques. Segmentation and supervised classification techniques will be applied to multi-frequency backscatter data to discriminate and map the broad landscape units, glacial boundaries and land cover classes (including rocky plains, bare ground, lakes and wetlands). Existing coastal polygons, field survey and optical data will be acquired through the AAD to assist in interpretation of SAR data and validation of classification outcomes. Data requirements: A series of multi-date, dual polarisation, StripMAP images acquired by TerraSAR-X over 2009 will be required for the project. At least 2 pairs are needed for each of the three islands for DEM generation and terrain characterisation. We may evaluate the potential of TerraSAR-X data acquired in SpotLight mode as well. The project has funds of $20,000 available for use in 2010. Deliverables: Updated DEMs will be generated for Macquarie, Heard and McDonald Islands at X- and L-bands. The DEM for McDonald Island will be the first of its kind. The analysis of topographic change by DEM differencing will provide new data for use in environmental studies. New spatial map data on the landforms and vegetation units will also be generated. Available optical data (including Quickbird imagery) will be draped over the DEMs for 3D reconstructions.

Back to list of proposals

DLR 2004-2016