基于l_k范数正则化的SAR图像目标特征增强  

Target Feature-Enhanced of SAR Image Based on Regularization of l_k Norm

作　　者：汪雄良[1] 王正明[1] 

机构地区：[1]国防科技大学理学院数学与系统科学系,长沙410073

出　　处：《电子与信息学报》2006年第9期1594-1597,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60272013);2001年全国优秀博士论文作者专项基金项目(200140)资助课题

摘　　要：增强SAR图像的目标特征对自动目标识别等具有重要意义。该文改进了一种基于lk范数正则化方法,并用于SAR图像目标特征增强。该方法通过开发利用符合SAR图像统计特性的先验知识,构造附加约束,把图像目标特征增强问题规划为形式简单的最优化问题,并利用一种迭代算法进行快速求解。仿真和实测数据计算结果证实了该方法的有效性。Target feature-enhanced processing of SAR image is meaningful to SAR ATR. One regularization method based on lk norm used for target feature-enhanced of SAR image is discussed in this paper. This method exploits the useful prior information which is well consistent to the statistically property of SAR image, makes up the additional constraint condition, turns the problem of target feature-enhanced processing of SAR image into the simple-formed optimization problem. A fast iterative algorithm is proposed to solve the optimization problem. The Simulation results and computational results of measured data prove its validity.

关 键 词：SAR图像 特征增强 LK范数 正则化 

分 类 号：TN957.52[电子电信—信号与信息处理]