Terrain Slope Correction and Precise Registration of SAR Data for Forest Mapping and Monitoring

SAR data acquired over hilly terrain show geometric and radiometric distortions due to the sidelooking configuration of the radar sensors. These effects usually lead to a distortion of the useful backscatter information related to land cover or bio-geophysical parameters. Post-processing approaches to remove such distortions are very important to broaden the application possibilities of radar remote sensing for land applications, especially in areas with significant terrain variation as can be found in many parts of Tasmania. Accurate orthorectification and terrain illumination correction for optical imagery have been well addressed in previous investigations. In this paper we present a new adaptation of algorithms which use the cosine of the surface tilt angles to more accurately correct for terrain slope illumination variations. We demonstrate the operational use of our approach using ALOS-PALSAR imagery. Critical to this task, and monitoring more generally, is accurate coregistration of data. To this end we provide a co-registration accuracy assessment of SAR data (including ALOSPALSAR and RADARSAT-2 data) and compare it with Landsat imagery sourced from an operational forest monitoring system. Keyword: geocoding, radiometric calibration, terrain slope correction, backscatter, forest mapping.