基于二维非参数特征分析的SAR图像目标识别  

SAR Image Target Recognition Based on Two Dimensional Nonparametric Feature Analysis

作　　者：刘振[1] 姜晖[1] 徐海峰[1] 

机构地区：[1]解放军电子工程学院,合肥230037

出　　处：《电讯技术》2012年第12期1940-1945,共6页Telecommunication Engineering

摘　　要：在分析传统Fisher线性鉴别分析局限性的基础上,由图像的行信息和列信息提出了两种形式的二维非参数特征分析(2DNFA)的特征提取方法,并应用于SAR图像目标的识别。直接在SAR图像矩阵上使用非参数特征分析提取特征不仅能充分发挥非参数特征分析的性能而且保留了图像矩阵的结构信息,大大降低了散度矩阵的维数,减小了运算量。使用美国MSTAR计划录取的数据对算法进行了仿真验证,实验结果显示两种形式的二维非参数特征分析在较低特征维数下的识别率均可以达到98%以上,表明所提方法的有效性和正确性。Based on the analysis of traditional Fisher Linear Discriminant Analysis（FLDA）, two forms of feature extraction methods of two dimensional Non_parametric Feature Analysis（NFA） based on the image＇s row informa- tion and column information are proposed and also applied to SAR （Synthetic Aperture Radar） image target recognition. The method that extracts feature by using non-parametric feature analysis directly on the SAR image matrix can not only give full play to the performance of nonparametric feature analysis but also retain the struc- tural information of the image matrix, and greatly reduce the dimension of the scatter matrix and the computing complexity. Experiments on MSTAR （Moving and Stationary Target Acquisition and Recognition） public database show that recognition rates of the new methods can achieve more than 98 % with less feature dimension- ality, indicating the correctness and effectiveness of the proposed methods.

关 键 词：合成孔径雷达 目标识别 FISHER线性鉴别分析 非参数特征分析 

分 类 号：TN957.52[电子电信—信号与信息处理] TP751[电子电信—信息与通信工程]