摘要
本文针对毫米波雷达距离像识别这一应用背景,提出了一种应用极化信息使距离像幅值趋于平稳,用神经学习来提取距离像角不变特征的模式识别新方法;其特点是通过极化处理减弱斑纹可增强对目标的确认,经神经学习获取的子类型特征与常规方法相比载有更多信息因而识别率高;实验结果证明了该方法的有效性.
In this paper,we propose a new pattern recognition approach using polarimetric information to make range profile amplitudes stable,and extract angle invariant features of range profiles using neural learning for the application of millimeter wave (MMW) radar range profile recognition. Its features are that the speckle-reduction by polarimetric processing may improve determination of the targets. By comparing with the conventional approach subclass features obtained by the neural learning,the new approach has carried more information and thus makes the rate of target recognition higher. The simulating resultes have approved the validity of the approach.
出处
《系统工程与电子技术》
EI
CSCD
1997年第7期27-30,共4页
Systems Engineering and Electronics
关键词
毫米波雷达
目标识别
成像系统
雷达
Radar target recognition,Range profiles,Polarimetric processing.
