A new radar target classification approach based on polarimetric high range resolution

Author

Yicheng, Jiang ; Yongtan, Liu ; Ping, Yu

Author_Institution

Res. Inst. of Electron. Eng., Harbin Inst. of Technol., China

fYear

1996

fDate

8-10 Oct 1996

Firstpage

147

Lastpage

150

Abstract

A new millimeter-wave (MMW) radar target classification approach has been proposed using polarimetric information to obtain stable amplitudes of range profiles, and neural learning to extract angle invariant features of range profiles. The means of the polarimetric processing for reducing the speckle can enhance ability to discriminate targets. Compared with conventional approaches, the subclass features obtained carry more information due to the neural learning and thus make the correctness of target classification higher. The simulation results have verified the validity of this approach

Keywords

feature extraction; learning (artificial intelligence); radar cross-sections; radar polarimetry; radar signal processing; radar target recognition; self-organising feature maps; speckle; vector quantisation; Kohonen model; angle invariant features; feature extraction; millimeter wave radar; modified learning vector quantization; neural learning; polarimetric high range resolution; polarimetric information; polarimetric processing; radar target classification; range profiles; scattering centers; self organising feature maps; simulation results; speckle reduction; stable amplitudes; subclass features; target discrimination; Feature extraction; Fingerprint recognition; Interference; Noise level; Noise reduction; Polarization; Radar imaging; Radar polarimetry; Speckle; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Radar, 1996. Proceedings., CIE International Conference of

Conference_Location

Beijing

Print_ISBN

0-7803-2914-7

Type

conf

DOI

10.1109/ICR.1996.573793

Filename

573793