海面油膜厚度SAR遥感探测机理及反演方法研究

Owing to no theoretical breakthrough, studies on detection mechanism and retrieval method of oil spill thickness are important and difficult technology. Based on Bragg scattering theory, a polarization ratio model of air-oil spill film-sea water medium will be studied in this project. Appling Fresnel reflection coefficients of multilayer medium into studying on oil spill sea surface, new concept of effective relative dielectric constant of oil film-sea water medium will be presented. According to relative dielectric constants of air and sea water, a relative dielectric constants retrieval model of oil spill film will be developed. Full polarization SAR images are used to validate this model. Relative dielectric constant models whose surface are full of and part full of oil spill film will be developed and used to theoretical studying on physical mechanisms of discrimination and thickness retrieval of oil spill film. Exploring spatial and time quasi-synchronization ocean color images (such as Landsat TM5/7) will be used to match SAR images. According to “BONN Agreement” Oil appearance code, ocean color images are used to validate results of oil spill thickness retrieval by SAR. Finally, a retrieval method of oil spill thickness will be concluded.

由于相关理论尚未有所突破，海面油膜厚度SAR遥感探测机理及反演方法是当前研究难点和技术瓶颈。本项目拟以Bragg全极化电磁散射理论为基础，建立大气-油膜-海水分层介质构成的油膜覆盖海面极化比模型，将分层介质菲涅耳反射系数模型应用于溢油海面情形，提出油膜-海水介质等效相对介电常数的概念。根据具体的大气和海水相对介电常数，建立油膜相对介电常数反演模型，利用海面溢油全极化SAR图像进行验证与模型修正。建立分辨单元内油膜全覆盖相对介电常数模型和油膜+海水覆盖相对介电常数模型，以此研究海面油膜识别物理机理和海面油膜厚度SAR遥感探测机理。结合时空准同步水色影像（如：Landsat TM5/7），判定和筛选出多种海面溢油SAR图像，根据《波恩协议》溢油外观识别码，对海面油膜厚度SAR遥感探测机理理论结果进行验证。最后提出海面油膜厚度SAR遥感反演方法。

本项目主要开展了溢油识别、油膜厚度和海面风反演方法研究，获得的如下研究成果：建立海面油膜电磁散射模型，揭示了SAR极化比探测海面油膜机理;提出了基于SAR极化比和纹理特征和基于全极化SAR图像的人工智能自动提取溢油特征的海面溢油识别方法, 揭示了低风速海面区域域海面油膜识别机理；提出了SAR风场反演新方法，获得了不同SAR图像分辨率下海面风场反演结果，为研究溢油厚度反演受风场和分辨率影响提供了理论支撑；提出一种消除低风速海面溢油误判的SAR溢油识别方法；以Bragg全极化电磁散射理论为基础，建立大气-油膜-海水分层介质构成的油膜覆盖海面极化比模型，利用海面溢油全极化SAR图像进行验证与模型修正；以此研究了海面油膜识别物理机理和海面油膜厚度SAR遥感探测机理与方法。项目结题时发了表论文10篇，其中SCIE论文7篇；培养博士生一名，硕士生3名，超额完成立项指标要求。