基于小波分解与方位角平均HRRP的SVM目标识别方法  

SVM Target Identification Method Based on Wavelet Decomposition and Azimuth Average HRRP

作　　者：梁海涛[1,2] 张学礼[1] 童创明[1] 王晓丹[1] 

机构地区：[1]空军工程大学导弹学院 [2]中国人民解放军93861部队

出　　处：《数据采集与处理》2010年第1期29-32,共4页Journal of Data Acquisition and Processing

基　　金：陕西省自然科学研究计划(2005F23)资助项目

摘　　要：针对雷达目标高分辨距离像(HRRP)方位角敏感以及以此为特征时正确分类识别率不高的问题,提出一种小波分解与按一定范围方位角平均的综合处理方法,降低了HRRP的维数,减小了HRRP的方位敏感性,增强了HRRP的稳定性,提取了雷达目标的主要特征;研究了结合支持向量机(SVM)的雷达目标分类识别方法,并对两种飞机目标缩比模型的原始HRRP数据、小波分解与方位角平均综合处理HRRP数据进行了分类识别实验,结果表明,小波分解、方位角平均以及SVM相结合的方法能够显著提高雷达目标的正确分类识别率,且稳定性更高,证明了方法的有效性。Aimed at the problems of the sensitive azimuth of high resolution range profile （HRRP） and lower classifying and identifying accuracies a method for wavelet decomposition and azimuth averaging in a certain range is proposed. By using the method, the dimension and the azimuth sensitivity of HRRP are reduced, and the stability of HRRP is improved. The main target characteristic is extracted. The method for support vector machine （SVM） is studied and adopted to classify and identify radar targets. The original HRRP data, wavelet decompo- sition and HRRP data obtained by azimuth averaging are used to classify and identify two dif- ferent scaling airplane models. The result indicates that the method can improve the classifying and identifying accuracies and enhance its stability. The effectiveness of the method is proved.

关 键 词：目标识别 小波分解 高分辨距离像 支持向量机 

分 类 号：TN95[电子电信—信号与信息处理] TP391[电子电信—信息与通信工程]