End-to-End Classification of Radar Target under Low SNR via Two-stream Fusion Network

Jiayin Xue, Fanyang Meng, Rongxiang He, Qinyu Zhang

Abstract

At present, most target recognition and classification approaches are based on inverse synthetic aperture radar (ISAR) imaging and feature extraction of the target image. Consequently, the discriminating performance severely depends on the imaging quality. Especially in the poor signal-to-noise ratio (SNR) environment, since the signal is usually immersed in strong noise, the image quality obtained by the conventional imaging method is extremely poor, which is not conducive to feature extraction and target recognition. To solve this problem, an end-to-end identification approach based on the neural network is proposed in this paper. Aiming at the characteristic of the echo data, a two-stream fusion convolutional neural network (CNN) architecture is designed to realize the end-to-end processing of “input data-output discriminant”. The simulation results verify the effectiveness of the proposed approach. In particular, it presents better robustness under low SNR.
BibTeX
@inproceedings{xue2019end,
  author = {Jiayin Xue and Fanyang Meng and Rongxiang He and Qinyu Zhang},
  title = {{End-to-End Classification of Radar Target under Low SNR via Two-stream Fusion Network}},
  booktitle = {WCSP},
  year = {2019},
  doi = {10.1109/wcsp.2019.8927968},
}

Loading related content...

Figure