Three-dimensional reconstruction of urban areas by multi-aspect TomoSAR data fusion

Abstract

Recent advances in airborne synthetic aperture radar tomography (TomoSAR) enable the generation of three-dimensional point clouds of urban areas with sub-meter accuracy and a point density comparable to LiDAR by exploiting multi-baseline stacks from multiple viewing angles. This paper summarizes the complete workflow from tomographic height reconstruction via geocoding to the combination of multi-aspect point clouds by a novel voxel-space-based fusion strategy. The evaluation results with respect to LiDAR-derived reference data show the great potential of multi-aspect TomoSAR data fusion when it comes to accurate and comprehensive reconstruction of urban areas.