基于压缩感知的合成孔径雷达图像目标识别  

SAR IMAGES TARGET RECOGNITION BASED ON COMPRESSED SENSING

作　　者：季秀霞[1] 卞晓晓[2] 

机构地区：[1]南京航空航天大学电子信息工程学院,江苏南京210016 [2]南京航空航天大学金城学院,江苏南京211156

出　　处：《计算机应用与软件》2013年第12期120-123,共4页Computer Applications and Software

基　　金：航空科学基金项目(2011ZC52034)

摘　　要：基于合成孔径雷达图像的目标识别技术在军事、民用等领域都具有十分重要的作用。针对SAR(Synthetic Aperture Radar)图像在像素域稀疏表示识别算法中存在的高维问题,在分析其图像统计特性的基础上,提出一种基于压缩感知的合成孔径雷达图像目标识别方法。该方法利用扩展最大平均相关高度滤波器训练样本,生成模板,提取模板广义二维主分量特征构成过完备字典,求解测试样本在字典下的稀疏表示系数,根据系数能量特征完成分类识别。对MSTAR数据库中合成孔径雷达图像进行仿真实验,结果表明,该方法复杂度低,识别时间短,是一种可行且有效的合成孔径雷达图像目标识别方法。Target recognition in SAR images plays an important role in military, civil and other areas. A method based on compressed sensing is presented for SAR target recognition after analysing the statistical characteristic of SAR images in order to solve the high dimensional problem of SAR image in pixel domain with sparse representation recognition algorithm. The method trains the samples and generates templates using the extended maximum average correlation height filter, extracts the template＇ s generalised two-dimensional principal component features to form an over-complete dictionary, the sparse representation coefficient of the test sample＇ s feature is computed base on the optimal dictionary. Classification and recognition are realised according to the energy feature of coefficient. Simulation experiment is carried out based on SAR images in MSTAR database, results show that the proposed method has lower complexity and short recognition time, it is a feasible and effective method for SAR images target recognition.

关 键 词：压缩感知 稀疏表示 合成孔径雷达 目标识别 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] TN957.52[电子电信—信号与信息处理]