摘要
本文以信号处理为基础，研究了识别特征的方法，包括直接利用距离像作为识别特征及提取距离像的频域特征和自相关函数特征，以避免距离时延对识别的不利影响并获得更多的高价统计信息；结合傅立叶变换的时移不变性和梅林变换的尺度不变性，讨论了距离像的傅立叶－梅林变换在改善目标特征稳定性方面的作用；研究了一种修正的直接梅林变换（ＭＤ－ＭＴ）算法。对几种提取的特征均采用相关滤波法对高分辨率的雷达目标进行识别。
Based on the signal processing the paper study up on the effective methods of target identifica-tion. In this paper, several means of extracting features are discussed here, including using range profile directly and extracting auto - correlation function to get more information of high order statistics and eliminate the effect of range shift. The shift invariance property of Fourier Transform and scale invariance property of Mellin Transform to extract feature with more stability, and a modified direct Mellin Transform (MDMT)algorithm are developed. For several extracted features, high range resolution radar targets are all identified with correlation filering. Therefore, the study of the methods of the HRR radar target identification is feasible and is valuable in engineering application.
出处
《指挥技术学院学报》
1999年第3期14-19,共6页
Journal of Institute of Command and Technology
关键词
雷达
目标识别
高距离分辨率
信号处理
radar target identification
high range resolution radar
radar range profile
feature extrac- tion
Fourier - Mellin transform
