摘要
在雷达目标识别中，由于雷达目标一维像特征提取容易，而且具有较好的识别效果，因此近来逐渐得到应用．但由于它是以散射点分布模型为依据的，对姿态变化非常敏感，为了提高对姿态变化范围较大的目标一维距离像的识别能力，在本文中，我们基于子波变换的多分辨分析理论及树形分层理论，提出一种级联SOM网络结构用于高分辨一维距离像的识别，并获得了令人满意的结果．
The 1-D range profiles can be extracted easily and have better identification result. They have been frequently used as identification feature in radar target identification recently. However, the 1-D range profiles of targets are based on the model of scattering centers, they have strong aspect angle dependence. In order to improve the identification performance of the radar target identification system based on 1-D range profile, a tree structure cascade SOM network, which is on the basis of multiresolution analysis theory, is proposed. And some satisfactory identification results are achieved.
出处
《北京工业大学学报》
CAS
CSCD
1998年第4期17-24,共8页
Journal of Beijing University of Technology
关键词
目标识别
一维距离像
雷达
SOM网络
子波变换
target identification, 1-D range profiles, neural networks, self-organizing map,wavelet transforms, multiresolution analysis
