The cross-scattering component of polarimetric SAR in urban areasand its application to model-based scattering decomposition
(English)In: International Journal of Remote Sensing, ISSN 0143-1161, E-ISSN 1366-5901Article in journal (Refereed) Accepted
After the work of Freeman, Durden, Pottier, and Yamaguchi, manydecomposition techniques have been proposed for urban areas, mainly to resolvethe overestimation problem of volume scattering. Since it has been validated thatthe cross-polarised (HV) scattering is caused not only by forests but also byrotated dihedrals, in this paper, we propose a cross scattering coherency matrix tomodel the HV component from orientated and complex buildings and thendemonstrate its performance on model-based scattering decomposition. Thebuilding orientation angle is considered in this coherency matrix, making itflexible and adaptive in the decomposition. Therefore, the HV components fromforests and orientated urban areas can be modelled respectively. Twodecomposition procedures are applied in this paper. The first one is to validatethe effectiveness of this scattering model. We regard the HV component fromurban areas as cross scattering, which is an independent scattering componentadded to the Yamaguchi four-component decomposition. Another one is theurban area decomposition application using this scattering model. Decompositionis implemented for urban and natural areas respectively and the HV componentfrom urban areas is regarded as their volume scattering. This procedure is similarto many other state-of-the-art methods for urban areas and needs to discriminatethe urban and natural areas before decomposition. Spaceborne Radarsat-2 C band,the Airborne Synthetic Aperture Radar (AIRSAR) L band and UninhabitedAerial Vehicle Synthetic Aperture Radar (UAVSAR) L band full polarimetricSAR data are used to validate the performance of this cross scattering coherencymatrix. The HV component of orientated buildings is generated, leading to abetter decomposition result for urban areas.
Place, publisher, year, edition, pages
Taylor & Francis Group.
cross scattering coherency matrix; orientated urban areas; modelbased decomposition; polarimetric SAR (PolSAR)
IdentifiersURN: urn:nbn:se:kth:diva-187939OAI: oai:DiVA.org:kth-187939DiVA: diva2:932588
QC 201606072016-06-012016-06-012016-06-07Bibliographically approved