基于压缩感知的综合孔径微波辐射测量射频干扰源定位方法研究

The practicability of remote sensing data from space-borne is seriously degraded by radio-frequency interferences (RFIs), which have limited the application and development of microwave remote sensing such as the retrieve of sea surface salinity. An essential solution to eliminating the RFI influences is to locate these RFIs and shut them off mandatorily through negotiation of ITU and local management institutions. Then how to accurately locate RFIs becomes an important problem to be solved. In the applications of aperture synthesis microwave radiometry, the existing RFI localization methods have predicaments such as the lack of spatial resolution performance, significant tailing effect of strong sources, failure to identifying the number of interferences, and so on. Therefore, a research on high-accuracy RFI localization methods using compressive sensing is proposed in the project. Following the route of “model establishment → algorithm development → performance enhancement”, researches are conducted as follows: analyzing the sparsity of interferometric arrays and a-priori information of the observed scene, establishing the model of sparse signal representation for RFI source localization, locating RFIs with a high accuracy by developing sparse reconstruction algorithms, investigating multiple-snapshot data processing approach to further enhance the localization performance, and verifying the feasibility and effectiveness of the proposed method through satellite-data based experiments. The results of this project can provide accuracy geolocation information for monitoring the illegal RFI sources. Meanwhile, these outcomes can be also applied in data analysis and processing from other satellites (such as “Gaofen-3” and “FY-4”), which will be helpful for improving the quality of information processing and then contribute to the application and development of space-borne remote sensing systems.

射频干扰（RFI）会严重影响星载遥感数据实用性，制约海表盐度反演等微波遥感应用与发展。消除RFI影响的根本途径是对RFI源精确定位后，通过国际电联与地方管理单位协调将其关闭，如何实现RFI源高精度定位是当前亟待解决的问题。在综合孔径微波辐射测量应用中，现有方法面临空间分辨率不高、强源拖尾效应明显、干扰源个数分辨不清等窘境，因此本项目基于压缩感知研究RFI源高精度定位方法，按照“模型构建→算法开发→性能提升”的思路，分析干涉式阵列结构稀疏特性与观测场景先验信息，构建RFI源定位的稀疏信号表征模型，再利用稀疏重构算法对RFI源进行高精度定位，并研究多快拍数据处理方法进一步提升定位性能，最后通过卫星实测数据验证方法可行性和有效性。本项目研究成果能为非法RFI监管提供精确位置信息，同时也有望用于“高分3号”、“风云4号”等卫星数据的分析与处理中，提高其信息处理质量，从而促进星载遥感系统应用与发展。

射频干扰（RFI）严重影响海洋盐度、土壤湿度探测卫星遥感精度，制约微波遥感应用与发展。面对干涉微波辐射测量迫切的应用需求，如何有效实现RFI检测与定位是当前亟待解决的痛点问题。本项目发展基于压缩感知的干涉辐射计RFI源定位方法，开展干涉式阵列扩展与信号模型、基于稀疏重构的RFI源定位算法、实测卫星数据的多快拍处理等理论方法研究。经过三年研究，完成了项目计划书中预定的全部研究目标，提出了基于非负复信号稀疏重构的综合孔径辐射计RFI高分辨鲁棒定位方法、基于约束矩阵完备化的综合孔径辐射计RFI高分辨定位方法、基于混合Coarray的镜像干涉综合孔径辐射计高性能定位方法、基于邻域重加权L1范数凸优化的综合孔径辐射计RFI高精度检测方法、基于重加权核范数凸优化的综合孔径辐射计RFI高精度检测定位方法、基于多拍空时联合稀疏恢复的综合孔径辐射计RFI高分辨定位方法等多种先进的处理方法，经过仿真与实测数据试验验证，显著提高了综合孔径辐射计RFI检测定位能力。以第一作者/通讯作者发表了论文23篇（SCI收录15篇和EI收录8篇），包括IEEE Trans. GRS/AP/MTT（11篇）、Information Fusion（1篇）、IEEE GRS/AWP Lett.（2篇）、Remote Sensing（1篇）等，获IEEE ICSIDP 2019国际学术会议优秀论文奖，申请国家发明专利6项。项目负责人入选了湖北省重大人才工程“楚天学者计划”和武汉市人才计划“武汉英才（优秀青年人才）”，协助团队培养了3名博士研究生和1名本科特优生。本项目研究成果可直接应用于微波遥感，还能迁移到辐射计目标成像探测应用中，对军事和民用领域具有重要意义。