组合字典下超宽带穿墙雷达自适应稀疏成像方法

发布时间：2018-10-05 20:29

【摘要】：针对现有超宽带穿墙雷达稀疏成像算法大多只采用点目标稀疏基表示模型和稀疏优化的正则化参数不能被自适应调整以及目标位置不在划分网格上带来虚假像的问题,该文提出一种基于贝叶斯证据框架的自适应稀疏成像方法。该方法首先利用组合字典独立稀疏表示场景中的点目标和扩展目标,然后在建立的偏离网格稀疏表示模型的基础上分层最大化各参数的似然函数,用第1层推理结合共轭梯度算法估计组合字典的各稀疏表示系数,用第2层推理估计正则化参数和目标的偏离网格量,最终通过迭代优化参数的设置得到问题的求解。仿真和实验结果表明,该方法不仅同时自适应增强穿墙场景中的点目标和扩展目标,还消除了偏离网格目标引起的虚假像。
[Abstract]:In order to solve the problem that most of the existing thinning imaging algorithms only use the sparse basis representation model of point target and the regularization parameters of sparse optimization can not be adaptively adjusted and the location of the target does not bring false image on the mesh. An adaptive sparse imaging method based on Bayesian evidence framework is proposed in this paper. In this method, the point targets and extended targets in the scene are represented by combinatorial dictionary independently, and then the maximum likelihood function of each parameter is maximized on the basis of the established sparse representation model. The first level reasoning combined with conjugate gradient algorithm is used to estimate the sparse representation coefficients of the combinatorial dictionary, and the second level reasoning is used to estimate the regularization parameters and the deviation of the target from the grid. Finally, the solution of the problem is obtained by iterative optimization parameter setting. The simulation and experimental results show that the proposed method not only adaptively enhances the point target and the extended target in the scene of penetrating wall, but also eliminates the false image caused by the deviation from the grid target.
【作者单位】：广西无线宽带通信与信号处理重点实验室;桂林电子科技大学信息与通信学院;
【基金】：国家自然科学基金(61461012) 广西区自然科学基金(2013GXNSFAA019329;2013GXNSFAA019004) 认知无线电与信息处理教育部重点实验室2015主任基金项目(CRKL150107)~~
【分类号】：TN957.52

【参考文献】