项目名称: 基于缺失数据分析和信息几何理论的SAR图像自动目标识别研究
项目编号: No.61501152
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 无线电电子学、电信技术
项目作者: 杨萌
作者单位: 杭州电子科技大学
项目金额: 19万元
中文摘要: 针对SAR ATR系统中目标特征信息损失或数据缺失、数据库所需目标特征信号内容数量庞大等问题,本项目开展基于缺失数据回归分析和信息几何理论的SAR图像自动目标识别研究。具体研究内容有:.一、构建缺失数据回归模型,包括. 1) 研究基于电磁散射特性的目标图像回归模型;. 2) SAR目标特征信息损失或有效数据缺失的回归分析;.二、研究统计流形的几何结构,包括 . 1) 研究参数空间的统计流形构建及其几何结构;. 2) 研究模型参数特征值的敏感性分析方法;.三、研究流形空间中的统计分类方法,包括. 1) 研究目标关键属性或特征值的概率量化方法;. 2) 研究流形空间中的目标类型评估方法。
中文关键词: SAR目标识别;缺失数据;信息几何
英文摘要: This application aims at missing data, large computation and other issues in SAR automatic target recognition (ATR). ATR in SAR image is chosen as study object. The project focuses on SAR ATR system via missing data analysis and information geometry. Firstly, missing data model for SAR image is constructed: 1) Regression model based on electromagnetic scattering theory, and its statistical characterization are studied; 2) Missing data analysis issues are discussed; Secondly, geometric structure of statistical manifold in parameter domain is constructed: 1) Properties of statistical manifold for target image data are studied; 2) Uncertainty and sensitivity analysis techniques are used in performance assessment for parametric analysis. Thirdly, statistical decision-making models in the manifold space are designed: 1) Probabilistic quantification methods of characteristic parameters for target are studied; 2) Statistical decision-making methods in the manifold space are designed.
英文关键词: SAR ATR;missing data;information geometry