红外单帧弱小目标检测算法研究综述  

A review of infrared single-frame dim small target detection algorithms

作　　者：刘征 杨德振 李江勇[1] 黄成章 LIU Zheng;YANG De-zhen;LI Jiang-yong;HUANG Cheng-zhang(North China Research Institute of Electro-Optics,Beijing 100015,China)

机构地区：[1]华北光电技术研究所,北京100015

出　　处：《激光与红外》2022年第2期154-162,共9页Laser & Infrared

摘　　要：随着红外预警探测技术的发展以及军事化作战需求的提高,满足低虚警、高检测精度的红外弱小目标检测算法一直是国内外研究重点。本文首先从成像特点、数学模型构建以及背景杂波干扰等方面阐述了弱小目标的特征与检测难点。然后分类介绍了近些年来提出的单帧弱小目标检测算法,并对算法的优势和不足进行了分析。最后结合当前红外预警探测领域的实际应用需求,分析了红外弱小目标检测算法未来的发展趋势。With the development of infrared early warning detection technology and the improvement of militarized combat requirements,the infrared dim small target detection algorithm with low false alarm and high detection accuracy has been the focus of research at home and abroad.Firstly,the characteristics and detection difficulties of dim small targets are expounded from the aspects of imaging characteristics,mathematical model construction,and background clutter interference.Afterward,the algorithms of single-frame dim small target detection proposed in recent years are introduced and classified,as well as the advantages and disadvantages of these algorithms.Finally,considering the practical application demands in the field of infrared early warning detection,the perspective of infrared dim small target detection algorithms is analyzed.

关 键 词：红外弱小目标 预警探测 阈值分割 神经网络 

分 类 号：TN976[电子电信—信号与信息处理] TP391.41[电子电信—信息与通信工程]