基于小波包变换和改进SVD的特征提取

Feature Extraction Based on Waveletpacket Transform and Modified SVD

提出了基于小波包变换和改进奇异值分解的高分辨雷达目标一维距离像特征提取方法。利用该方法对三种目标分类进行仿真实验,结果表明:该特征提取方法,对平移和姿态角变化有较强的稳健性,说明该算法是有效的。

A feature extraction method of high-range-resolution radar profiles, which takes advantage of wavelet packet transform and modified SVD(singular value decomposition) was proposed. We used this method to classify 3 different targets. The result manifests this method is robust and efficient.