基于稀疏表示的SAR图像目标识别方法  

SAR ATR method based on sparse representation

作　　者：刘振[1] 姜晖[1] 王粒宾[1] 

机构地区：[1]电子工程学院信息工程系,合肥230037

出　　处：《计算机工程与应用》2014年第10期212-215,232,共5页Computer Engineering and Applications

摘　　要：为了准确地进行SAR图像目标识别,提出一种基于稀疏表示的SAR目标识别方法,在用主成分分析(PCA)进行降维的前提下,利用降维后的训练样本构建稀疏线性模型,通过?1范数最优化求解测试样本的稀疏系数解x,利用系数的稀疏性分布进行目标的分类识别。基于MSTAR数据进行了仿真验证,实验证明,基于稀疏表示的SAR目标识别方法在一定的特征维数下能够获得很好的识别性能,在目标方位角未知的情况下识别率仍可达到98%以上。In order to recognize SAR target accurately, an identification method based on sparse representation is proposed. The training samples after dimensionality reduction using principal component analysis are used to build a sparse linear model. The sparse coefficient solution x of the test sample is solved by ？1 -minimization. The identification task is solved by utilizing the sparse distribution of the sparse coefficient. Experimental results with MSTAR dataset verify that the identification method based on sparse representation in a certain characteristic dimension can obtain good recognition performance, and the recognition rate can reach more than 98%without knowing the target azimuth.

关 键 词：合成孔径雷达(SAR) 目标识别 稀疏表示 ξ1范数最优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN951[自动化与计算机技术—计算机科学与技术]