Radar data processing, quality analysis and level-1b product generation for AGRISAR and EAGLE campaigns

Abstract

For the assessment of specific programmatic needs of the Sentinel program and for the understanding and quantification of bio-geophysical parameters of different vegetated surfaces ESA has initiated and partially funded two airborne campaigns in 2006. In both campaigns the airborne SAR system of DLR, E-SAR was employed to acquire multi-frequency and multi-polarisation data over the different test sites. For the AGRISAR campaign 16 E-SAR flights have been conducted over Görmin test-site in the period from April-August 2006 to acquire representative data sets which cover the complete agricultural vegetation period of different crops. For the EAGLE project two test-sites have been flown, Cabauw for specific investigations on agricultural areas and Speulderbos for forest parameter assessment. The present paper describes the adopted methodology for processing the acquired E-SAR data. First the standard processing approach which leads to RGI (radar geometry images) and GTC (geocoded terrain corrected) products is described. Then we focus on the special aspects which were considered for the processing of the temporal time series for the AGRISAR project. Further discussion is included with respect to the mosaicked DEM, which was generated for the AGRISAR test site from E-SAR single-pass SAR interferometry, as well as the generation of synthetic QUAD-POL products in C-band (performed for both AGRISAR and EAGLE campaigns). Another section is devoted to the generation of Sentinel-1 like data quality products from high resolution C-band data, as is requested by the specific programmatic needs of ESA. The paper concludes with information on the data accessibility via the EOWEB interface of DLR.