自动目标识别的工程视角述评  

Automatic Target Recognition from an Engineering Perspective

作　　者：郁文贤[1] YU Wenxian(School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《雷达学报（中英文）》2022年第5期737-752,共16页Journal of Radars

基　　金：国家自然科学基金重点项目(61331015)。

摘　　要：自动目标识别(ATR)是一个和信号与信息处理、模式识别、人工智能等学科密切相关的特殊工程技术应用领域。由于ATR系统识别对象固有的不确定性,识别环境的复杂性,以及日益加剧的识别对抗性,使得ATR的发展一直面临着从理论到技术、到应用的系统性挑战。该文从工程视角出发,阐述了ATR的定义与内涵,简要回顾与分析了该领域发展动态,梳理了ATR的核心技术体系与系统开发模式,最后对未来的发展挑战做了分析。Automatic Target Recognition(ATR)is a special engineering application field which is closely related to signal and information processing,pattern recognition,artificial intelligence and other disciplines.Owing to the inherent uncertainty of ATR systems,the complexity of the recognition environment,and the increasingly adversarial nature of recognition,the development of ATR faces systematic challenges from theory to technology and applications.This paper presents the definition and connotation of ATR from an engineering perspective,briefly reviews and analyzes the developments in this field,explores the core technology system and system development model of ATR,and finally examines the future development challenges.

关 键 词：自动目标识别 环境复杂性 识别对抗性 技术体系 工程应用 

分 类 号：TN957[电子电信—信号与信息处理]