Exploitation of sublook tomography for the retrieval of 3-D point clouds in urban environments with high-resolution spaceborne SAR data

Abstract

Sublook tomography (SBT, Sublook SAR Tomography) is a recent technique that exploits the wide bandwidth of the signal transmitted by synthetic aperture radar (SAR, Synthetic Aperture Radar) to retrieve the vertical reflectivity profile of semitransparent media. In urban environments, although volumetric scattering generally does not occur, layover effects are frequent due to the geometric complexity of structures. Conventional tomographic approaches to resolve layover typically require multiple SAR acquisitions with diverse baselines. In contrast, the SBT technique is able to extract this information using only two images, significantly reducing the amount of data required. This work investigates the application of SBT to mitigate layover in urban environments using real data acquired from orbit. The limitations imposed by the reduced vertical resolution inherent to the method are analyzed, as well as its ability to identify and separate targets in height. To this end, the performance of SBT is evaluated in a spaceborne scenario using TerraSAR-X data over a region of Las Vegas, comparing two inversion methods: a classical spectral estimator and the SL1MMER algorithm, based on compressive sensing and suitable for sparse profile reconstruction. The comparison between SBT and conventional tomography allowed examining the capability of the SBT technique to estimate heights, recover multiple scatterers, and generate three-dimensional structural information in complex urban areas. Its practical limitations were also discussed, as well as its application potential and prospects for improvement.