StereoPol Radarsat-2 data fusion in radargrammetry

With new Radarsat-2 capabilities the complex polarimetric data can be included in the radargrammetry by fusing stereo polarimetric (stereopol) data into intensity stereo images. In this study the different fusion techniques, based on the combination of complex polarimetric data and polarimetric decompositions, are investigated. The stereo matching capability of the fusion techniques was quantitatively evaluated with textural parameters and cross-correlation coefficient and by comparing stereo-extracted digital surface models (DSM) with airborne lidar data. The textural parameters, calculated from gray level co-occurrence matrix (correlation and contrast), and the cross-correlation coefficient were found to be potential indicators of stereo matching of fused images. The application of the Pauli decomposition showed that stereo matching accuracy, evaluated by the difference between lidar and DSM, depends on the scattering mechanism.