贝叶斯模型下的OMP重构算法及应用  

OMP reconstruction algorithm via Bayesian model and its application

作　　者：李少东[1] 裴文炯[1] 杨军[1] 胡国旗[1] 

机构地区：[1]空军预警学院,湖北武汉430019

出　　处：《系统工程与电子技术》2015年第2期246-252,共7页Systems Engineering and Electronics

基　　金：军队重点项目资助课题

摘　　要：针对稀疏度先验信息缺失的条件下,正交匹配追踪(orthogonal matching pursuit,OMP)算法设置冗余稀疏度时,造成信号过重构、抗噪性能变差等问题,基于贝叶斯检验模型,提出了贝叶斯正交匹配追踪(Bayesian orthogonal matching pursuit,BOMP)算法。并推导了该算法估计信号的克拉美罗下界,最后将算法应用于逆合成孔径雷达(inverse synthetic aperture radar,ISAR)成像。理论分析和实验结果表明,由于该算法能够更加真实地估计信号支撑集,因而具有更好的重构精度、抗噪性能,同时降低了计算复杂度。Due to lack of sparsity information, the orthogonal matching pursuit （OMP） algorithm must set a redundant support in advance, which leads to heavier computation burden, lower noise suppressive abitity and poor signal reconstruction performance. This paper proposes a modified Bayesian orthogonal matching pursuit （BOMP）algorithm. Meanwhile, the Cramer-Rao lower bound （CRLB） of the estimated signal is analyzed. Fi- nally BOMP is used for inverse synthetic aperture radar （ISAR） imaging. Theoretical analysis and simulation results show that the proposed algorithm can improve the noise suppressive ability and reconstruction accuracy, and decrease the computing complexity obviously.

关 键 词：正交匹配追踪 贝叶斯模型 克拉美罗下界 逆合成孔径雷达成像 

分 类 号：TN951[电子电信—信号与信息处理]