雷达像智能识别对抗研究进展  

Intelligent Radar Image Recognition Countermeasures:A Review

作　　者：高勋章[1] 张志伟[1] 刘梅[1] 龚政辉 黎湘[1] GAO Xunzhang;ZHANG Zhiwei;LIU Mei;GONG Zhenghui;LI Xiang(College of Electronic Science and Technology,National University of Defense Technology,Changsha 410073,China)

机构地区：[1]国防科技大学电子科学学院,长沙410073

出　　处：《雷达学报（中英文）》2023年第4期696-712,共17页Journal of Radars

基　　金：国家自然科学基金(61921001)。

摘　　要：基于深度神经网络的雷达像智能识别技术已经成为雷达信息处理领域的前沿和热点。然而,深度神经网络模型易受到对抗攻击的威胁。攻击者可以在隐蔽的条件下误导智能目标识别模型做出错误预测,严重影响其识别精度和鲁棒性。该文梳理了近年来雷达像智能识别对抗技术发展现状,总结分析了现有雷达一维/二维像识别对抗攻击方法和防御方法的特点,最后讨论了当前雷达像智能识别对抗研究领域值得关注的5个开放问题。Intelligent radar image recognition based on Deep Neural Networks(DNN)has become an important topic in radar information processing.However,DNN models are susceptible to adversarial attacks.Malicious attackers can cause intelligent image recognition models to make incorrect predictions,considerably reducing their recognition accuracy and robustness.This article reviews recent research progress on intelligent radar image recognition countermeasures.Then it summarizes the adversarial attack methods on one/two-dimensional radar image recognition models and adversarial defense methods.Finally,it discusses five open questions worthy of in-depth research in intelligent radar image recognition countermeasures.

关 键 词：雷达像识别 深度神经网络 对抗攻击 后门攻击 对抗防御 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置]