计算机与现代化 ›› 2020, Vol. 0 ›› Issue (08): 14-20.doi: 10.3969/j.issn.1006-2475.2020.08.003

• 图像处理 • 上一篇    下一篇

基于谱归一化生成对抗网络的目标SAR图像仿真方法

  

  1. (1.中国科学院空天信息创新研究院,北京100094;2.中国科学院大学,北京100049;
    3.中国科学院电子学研究所,北京100190)

  • 出版日期:2020-08-17 发布日期:2020-08-17
  • 作者简介:孙智博(1994-),男,山西太原人,硕士研究生,研究方向:SAR图像处理,深度学习,E-mail: sunzhibo17@mails.ucas.ac.cn; 徐向辉(1974-),男,研究员,博士,研究方向:微波成像技术,雷达信号处理,E-mail: xhxu@mail.ie.ac.cn。
  • 基金资助:
    国家重点研发计划项目(2017YFB0503001)

Simulation Method of Target SAR Image Based on  Spectral Normalization Generative Adversarial Network

  1. (1. Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China;
    2. University of Chinese Academy of Sciences, Beijing 100049, China;
    3. Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China)
  • Online:2020-08-17 Published:2020-08-17

摘要: 为解决合成孔径雷达(Synthetic Aperture Radar, SAR)自动目标识别(Automatic Target Recognition, ATR)中的数据稀疏问题,提出一种基于谱归一化生成对抗网络(Spectral Normalization Generative Adversarial Network, SN-GAN)的目标SAR图像仿真方法。本文方法通过构建目标—场景—雷达耦合物理模型,求解散射强度分布图,利用SN-GAN实现对散射强度分布图的优化,生成高质量仿真SAR图像。通过3种相似性评估算法对仿真图像进行相似度评估,验证本文仿真方法的有效性。最后通过多组SAR ATR进行实验验证,在训练集中加入SN-GAN优化的仿真SAR图像可以有效缓解数据稀疏问题,提升分类算法的准确率。

关键词: SAR图像, 图像仿真, SN-GAN

Abstract: In order to solve the data sparse problem in Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), this paper proposes a simulation method of target SAR images based on SN-GAN (Spectral Normalization Generative Adversarial Network). The method obtains the scattering intensity distribution maps by constructing the coupled physical model among target, scene and radar, then refines the scattering intensity distribution maps by using SN-GAN to generate the high-quality simulated SAR images. The similarity evaluation of the simulated images is carried out by 3 kinds of similarity evaluation algorithms to verify the effectiveness of the simulation method. Finally, through multiple sets of SAR ATR experiments, it is verified that adding simulated SAR images optimized by SN-GAN to the training set can effectively alleviate the data sparse problem and improve the accuracy of the classification algorithms.

Key words: SAR image, image simulation, SN-GAN

中图分类号: