基于稀疏表示的SAR图像属性散射中心参数估计算法

发布时间：2018-07-04 14:26

本文选题：SAR + 属性散射中心　；参考：《电子与信息学报》2014年04期

【摘要】：考虑目标频率-方位2维观测数据在属性散射中心模型参数空间上的稀疏性,该文提出一种基于稀疏表示的属性散射中心提取与参数估计方法。由于模型参数维数较高,构造的高维联合字典将消耗较多系统资源。该算法通过分别构建包含位置信息与方位属性参数信息的两个低维字典代替高维的联合字典实现距离特性与方位特性的解耦合,以降低资源需求,并通过正交匹配追踪(OMP)-RELAX联合算法求解0l优化问题,从而实现在频率-方位角域上位置参数与方位属性参数的联合估计。根据提取的属性散射中心可以有效地估计目标或目标重要部件的几何尺寸。基于电磁计算数据和实测数据的实验结果验证了该算法的有效性。
[Abstract]:Considering the sparsity of target frequency-azimuth 2-D observation data in the parameter space of attribute scattering center model, this paper presents a method for extracting and estimating attribute scattering center based on sparse representation. Because of the high dimension of model parameters, the constructed high dimensional joint dictionary will consume more system resources. In this algorithm, two low-dimensional dictionaries containing location information and azimuth attribute parameter information are constructed instead of high-dimensional joint dictionaries to realize the decoupling of the distance and azimuth characteristics, so as to reduce the resource requirement. The orthogonal matching tracing (OMP) -RELAX joint algorithm is used to solve the 0l optimization problem, which realizes the joint estimation of position parameters and azimuth attribute parameters in frequency-azimuth domain. According to the extracted attribute scattering center, the geometric size of the target or its important components can be estimated effectively. Experimental results based on electromagnetic data and measured data verify the effectiveness of the algorithm.
【作者单位】：西安电子科技大学雷达信号处理国家重点实验室;
【基金】：国家自然科学基金(61271291,61271024,61201292,61201283) 新世纪优秀人才支持计划(NCET-09-0630) 全国优秀博士学位论文作者专项资金(FANEDD-201156) 中央高校基本科研业务费专项资金资助课题
【分类号】：TN957.52

【参考文献】