Compilation of a bi-temporal JERS SAR mosaic over the African rain forest belt in the GRFM project

The Global Rain Forest Mapping project GRFM is an international collaborative effort promoted by the National Space Development Agency of Japan, NASDA. The main goal of the project is to produce a wall to wall map of the entire tropical rain forest using the L-band SAR on board the JERS-1 spacecraft. Within the GRFM project the European Commission Joint Research Centre acts as main processing node for the assemblage and validation of the radar mosaics related to the African continent. The authors give an overview of the techniques used for the compilation of these wide area radar mosaics, with emphasis on the scene geo-referencing. A global optimization technique is used based on a least squares estimation of the scene geometry parameters given observations which comprise intra-scenes correlation measures, ground control points and nominal scene position derived by orbital parameters and the range-doppler equation. Two acquisitions at different dates are simultaneously included in the estimation process, thus assuring optimal co-registration between dates. A similar technique is used for radiometric calibration of the mosaic. Validation of the multi-temporal radar map's geometry with respect to existing cartographic data-a key issue in view of certain thematic applications-is also discussed. A RMSE (Residual Mean Squared Error) of 56 m was obtained when using tie-points between scenes only. This error characterizes the mosaic internal consistency for scenes down-sampled to a pixel spacing of 100 m. A RMSE of 240 m was obtained when using ground control points (GCP) derived from digital cartographic data.