第43卷第3期2023年6月弹箭与制导学报JournalofProjectiles,Rockets,MissilesandGuidanceVol.43No.3Jun.2023DOI:10.15892/j.cnki.djzdxb.2023.03.005收稿日期:2022-10-20作者简介:陈睿琦(1993—),女,工程师,硕士,研究方向:软件工程。引用本文:陈睿琦,南心蒙,常皓.距离信息与排序模型融合的显著性目标检测方法[J].弹箭与制导学报,2023,43(3):33-38.CHENRuiqi,NANXinmeng,CHANGHao.Asalientobjectdetectionmethodfusedwithdistanceinformationandmanifoldranking[J].JournalofProjec-tiles,Rockets,MissilesandGuidance,2023,43(3):33-38.距离信息与排序模型融合的显著性目标检测方法陈睿琦1,南心蒙1,常皓2(1西安现代控制技术研究所,陕西西安710065;232382部队,北京100072)摘要:针对图像探测系统对地面场景中目标探测及识别问题,提出基于距离轮廓信息与排序模型相融合的显著性目标检测方法,利用二维简易距离轮廓信息对可见光及红外图像实施分割,采用流形排序模型对图像进行处理,进而提高目标的检测性能。实验结果表明,白天条件下对显著性目标检测的准确率可达到87.7%,召回率达到91.6%;夜间条件下的检测准确率达到81.2%,召回率达到86.3%,证明了该方法在保证图像处理速度的前提下,能够显著提高目标检测的正确性。关键词:图像处理;显著性目标检测;目标识别;流形排序;距离信息中图分类号:E91;TJ765.3;TP391.4文献标志码:A文章编码:1673-9728(2023)03-0033-06ASalientObjectDetectionMethodFusedwithDistanceInformationandManifoldRankingCHENRuiqi1,NANXinmeng1,CHANGHao2(1Xi’anMordernControlTechnologyResearchInstitute,Xi’an710065,Shaanxi,China;2No.32382Unit,Beijing100072,China)Abstract:Aimingattheproblemofdetectingandrecognizingtargetsfromcomplexbackgrounds,asalientobjectdetectingmethodbasedonthecombinationofdistanceinformationandmanifoldrankingisproposed.Inordertoimprovethedetectionperformanceincomplexbackgrounds,two-dimensionaldistanceinformationisusedtosegmentvisiblelightimagesandinfraredimages,andmanifoldrankingisusedtoprocesstheimages.Experimentalresultsshowthatunderdaylightconditionthepreci-sionandrecallofthismethodare87.7%and91.6%respectively;whileundernighttimeconditiontheprecisionandrecallareabove81.2%an...