隐蔽伪装目标检测方法综述  

Review of Methods for Detecting Concealed Camouflage Targets

作　　者：张学凯 彭月平[1] 唐威 康文超 叶泽聪 ZHANG Xuekai;PENG Yueping;TANG Wei;KANG Wenchao;YE Zecong(School of Information Engineering,Engineering University of PAP,Xi'an 710086,China)

机构地区：[1]武警工程大学信息工程学院,陕西西安710086

出　　处：《陆军工程大学学报》2025年第1期61-70,共10页Journal of Army Engineering University of PLA

基　　金：国防科技创新自主选题研究项目(ZZKY20223105)。

摘　　要：伪装主要包括隐蔽、形变、干扰和降低显著性等。随着伪装技术的发展与进步,对隐蔽伪装目标检测的技术要求也逐步提高。针对隐蔽伪装,梳理了伪装技术的研究现状;论述了传统隐蔽伪装目标检测方法,特别是近年来基于深度学习的隐蔽伪装目标检测方法;进一步分析了隐蔽伪装目标检测常用的动物伪装数据集和军事迷彩数据集,并归纳了4个常用的评价指标;总结了隐蔽伪装目标检测在重叠遮挡、尺度变化、动态变化和设备资源受限等方面面临的挑战,并从3个不同角度探讨了隐蔽伪装目标检测在军事领域中的发展前景。Camouflage primarily encompasses concealment,deformation,interference and reduction of saliency,etc.With the advancement and progress of camouflage technology,the technical requirements for detecting concealed camouflage targets have gradually increased.Regarding concealed camouflage,this article reviews the current research status of camouflage technology;discusses the traditional methods for detecting concealed camouflage targets,especially those based on deep learning in recent years;further analyzes the commonly used datasets for concealed camouflage target detection,including animal camouflage datasets and military camouflage datasets,and summarizes four frequently employed evaluation metrics;summarizes the challenges faced by concealed camouflage target detection in aspects such as overlapping occlusion,scale variation,dynamic changes,and limited device resources;and explores the development prospects of concealed camouflage target detection in the military field from three different perspectives.

关 键 词：伪装技术 伪装目标检测 深度学习 零样本学习 

分 类 号：E951.4[军事—军事工程]