基于Hilbert-Huang变换的雷达信号特征提取技术

针对高分辨雷达目标散射回波的非平稳特性，提出了利用Hilbert—Huang变换进行时频分析获得目标特性的方法。首先利用简单电磁散射模型仿真了三类目标的一维距离像，然后采用Hilbert—Huang变换对各类目标在不同姿态角下的回波进行时频分析，提取时频分布图的边际谱作为目标特征矢量，利用径向基函数神经网络分类器对特征矢量进行训练和学习，最后对三类目标作了识别，仿真对比实验验证了该方法的有效性。

Feature Extraction of Radar Signals Based on Hilbert-Huang Transform

Using time-frequency analysis to the echoes of high-resolution radar scatting center can obtain the target characteristics.Hilbert-Huang transform has good characteristic of time-frequency cluster.This paper extracts the time-frequency characteristics of one-dimensional range profile of three different targets by Hilbert-Huang transform and then classifies the three different targets by RBF neural network classifier.The results illustrate that the method is effective.