3-D Joint Inversion of Urban Persistent and Distributed Scatterers via SAR Tomography

Abstract

Existing urban spaceborne synthetic aperture radar (SAR) tomography (TomoSAR) inversion methods predominantly focus on persistent scatterer (PS), while distributed scatterer (DS) constitute a substantial proportion of urban targets and provide critical complementary information for comprehensive urban inversion. To synergize PS and DS observations under limited data acquisition constraints, this study proposes a novel TomoSAR three-dimensional inversion framework for urban multiscattering targets. The framework establishes a three-tier network architecture: the reference network (RN), star network, and DS network. For RN construction, we propose a novel efficient and robust method, namely the directed minimum spanning tree RN. Meanwhile, to accommodate different practical requirements, two strategies for selecting construction targets are provided: one is to construct the RN using only PS, with DS appearing exclusively in the DS network; the other is to jointly construct the RN using PS and a subset of DS with relatively stable scattering properties and located in sparsely distributed PS regions. In the experimental validation, we present, for the first time, three-dimensional inversion results from joint PS and DS processing using 4-day repeat-pass LuTan-1 data over Shanghai, China.