Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal MTH0413

TitleFRA-SAR-2010: Application of Terra-SAR-X and multi parametric radar data to support the FAO Forest Resource Assessment 2010 (FRA-SAR 2010, Approved AO-project funding ref. FK 50EE0802).
Investigator Knuth, Ralf - Friedrich-Schiller-University of Jena, Department of Geography
Team Member
Dipl.-Geogr. Knuth, Ralf - Friedrich-Schiller-University of Jena, Department of Geography
Prof. Dr. Schmullius, Christiane - Friedrich-Schiller-University of Jena, Department of Geography
Dr. Thiel, Christian - Friedrich-Schiller-University of Jena, Department of Geography
Dr. Häusler, Thomas - GAF, N/A
Dr. Siegmund, Robert - GAF AG, N/A
N/A Wilkie, Mette - FAO Forestry Department, Forest Assessment and Reporting Service
N/A Gerrand, Adam - FAO Forestry Department, Forest Assessment and Reporting Service
SummaryThe overall objective of FRA-SAR 2010 is to develop a multistage analysis approach based on TerraSAR-X data to enhance national and trans-national forest inventories. The FRA-SAR project is funded by DLR and BMBF (FK 50EE0802). The project is part of the United Nations Global Forest Resource Assessment (FRA 2010) that employs for the first time a global Remote Sensing Survey of Forests (RSSF) in order to improve the national inventories and to further develop the data gathering methods of FAO. FRA-SAR 2010 is an integrated part of the survey to fill the potential data gaps due to limited availability of cloud-free optical data that are used as main data source for RSSF. FRA-SAR aims to test the application of TerraSAR-X data for the use in the global forest inventory and to develop a respective processing and analysis chain. Furthermore, the potential of multi-frequent radar composites to enhance forest type and forest structure mapping will be tested by integrating X-, C- and L-band data from TerraSAR-X, Enivsat and ALOS. A training handbook, describing the methodology will be prepared for the UN forest experts working in the frame of FRA 2010 and communicated in dedicated FAO training courses to allow for the use of SAR-data in locations with persistent cloud cover. Comparability and future application of the developed methodologies and products will be assured through reconcilement with other international programs (GMES and GEOSS) and mapping efforts (GlobCover) by the project partner GAF Company. FRA-SAR 2010 is therefore a project proposal for “Innovative methodologies with application potential and industry integration”. Reflecting the different objectives, the project is scheduled in several phases. First, the SAR data base will be established, including the required TerraSAR-X, Envisat ASAR and ALOS PALSAR data. Central part of the project is the method development of the processing and image analysis, comprising the method for mapping six forest classes required by FAO based on TerraSAR-X data and, in a second stage, the radar multi-frequency composite analysis for enhanced forest structure mapping. For the first part, different image processing techniques will be applied ranging from image interpretation and radar image filtering to sophisticated methods as texture analysis and image segmentation. The target is to map all six desired forest classes with at least 85 % overall accuracy. The study will be based on TerraSAR-X SPOT light data in a representative selection of 50 sampling sites. In a second development phase, methods will be tested to derive detailed forest structure parameters, in five selected test sites, where detailed in-situ data are required. The study will integrate all common analysis methods for combined X+C+L-band data, integrating the experience made with the data of the SIR-C/X-SAR shuttle mission. In the validation phase the developed methodology will be tested in independent validation sites. The next project phase employs the preparation of training materials for the FAO image analysis courses for the regional experts as these courses will include TerraSAR-X analysis for application in areas of persistent cloud cover. In total, 350 TerraSAR-X scenes in SPOT-light mode are needed for the project. A detailed acquisition plan will be developed after consultation with FAO, when the explicit data gaps within the optical imagery are known. After successful completion of the project, the several products and deliverables will be provided, including a multistage processing and analysis chain for mapping the FAO forest classes using TerraSAR-X, a methodology for the synergetic use of X-, L- and C-band data to derive forest structure information, validated products, training materials and scientific literature documenting the research progress.

Back to list of proposals

© DLR 2004-2016