Science Service System

Summary of Proposal COA1876

TitleA study on time-frequency and polarimetric methods for ship characterization
Investigator Tebaldini, Stefano - Politecnico di Milano, DEI
Team Member
Dr. Tebaldini, Stefano - Politecnico di Milano, DEIB
Prof Monti Guarnieri, Andrea - Politecnico di Milano, DEIB
eng. Banda, Francesco - Politecnico di milano, deib
Prof Ferro-Famil, Laurent - Université de Rennes 1, IETR
SummaryThe scope of the proposed work is todiscuss the added value of the exploitation of dual polarimetric Spotlight andHigh resolution Spotlight TSX data concerning ship detection, classification,and velocity estimation. The available data will be analyzed by investigatingthe variation of the target’s polarimetric signature with respect to squintangle, to the aim of revealing variations that can be related to the physicalnature of the scatterers and their motion during the acquisition time. Theanalysis will take into account target motion during the acquisition time, inorder to understand to what extent adaptive signal focusing and polarizationplane steering can be used to mitigate reflectivity variations over time andaspect angle. Results will be compared to those obtainedby considering i) single polarization data and ii) a limited squint angle span(i..e: comparable to strip-map), so as to highlight which benefits are broughtby the exploitation ofsuchhigh-resolution dually polarized TSX products. In addition, we expect this work to producea set of measurements useful to model the statics relative to the polarimetricvariation of different classes of vessels with respect to squint angle, to bedelivered to the Scientific Community in the form of a Journal paper (uponDLR’s approval, as specified in the Schedule Section) and to serve as areference for further studies in this field. Different approaches will be considered. Model-based approaches base on arepresentation of the SAR signal into a set of contributors. Each selectedcontributor is associated with a scatterer of the scene and is parameterized byrelevant physical descriptors, leading to a multidimensional model of theobject. This class of methods can be employed with both single ormulti-polarization data. A TimeFrequency signal model for time/frequency varying targets will be studied for the case of ships, using two statisticaldescriptors related to the signal stationary aspect and coherence in thetime-frequency domain, to the aim of improving detection and estimation for time/frequencyvarying targets . Finally, we will discuss the possibility toreuse concepts and methods from Polarimetric SAR Tomography, where polarizationand angular diversity are used to carry out target detection and localization withoutthe need to provide an a-priori parameterization of the targets. Data requirements: Data to be analyzed within this project arerequired to be Dual polarization Spotlight or Dual Polarization High ResolutionSpotlight TSX images acquired over maritime traffic routes, costal zones, andmain harbors.The data should beprocessed to L1 Single Look Slant range Complex (SSC) in radar coordinates(slant range – azimuth). A test search on the EOWEB site has shownthe availability of TSX products suitable to this research over the strict ofGibilter, and costal zones near Le Havre (France), Barcelona (Spain), and NewYork (USA). Deliverables: Progress reports shall include thefollowing information: activity progress and status, preliminary resultsobtained, variations to the plan (if any), updated investigators names (ifany). The final report shall be focused on the performed activities and on theresults obtained, focusing on the methodology description and on the comment ofthe results with respect to the objectives. The report shall be created in MSWord and made available to DLR in pdf format. It is foreseen to submit paperswith the results of the investigations to the international conferencesandworkshops (e.g. IGARSS, EUSAR, CEOS). The submission of a paper to a peer-reviewed international Journal shall becarried-out, the content of the paper to be evaluated and discussed beforehandwith DLR. Funding: The activities within the proposed workwill be self-funded by Politecnico di Milano and the University of Rennes I.

Back to list of proposals

© DLR 2004-2016