二维码（扫一下试试看！） | | 士兵目标的少样本深度学习检测方法 | | A Deep Learning Detection Method for Soldier Target Based on Few Samples | | 投稿时间：2020-12-09 | | DOI：10.15918/j.tbit1001-0645.2020.229 | | 中文关键词:兵器科学与技术深度学习目标检测目标跟踪 | | English Keywords:armament science and technologydeep learningtarget detectiontarget tracking | | 基金项目:国家部委基础科研计划资助项目（JCKY2019602C015） | | | 摘要点击次数:567 | | 全文下载次数:242 | | 中文摘要: | | 针对敌士兵数据集样本较少的问题，提出一种基于YOLOv3的少样本深度学习目标检测方法.利用数据增广提高少样本目标检测模型的鲁棒性，改进网络结构将浅层网络特征图跨层连接至深层网络，采用k-means聚类获取适合士兵目标特性的锚点框，利用预训练提高模型训练收敛速度.实验结果表明，本文方法对少样本敌士兵目标检测成功率mAP达到85.6%、检测精度IOU达到82.18%，且对小型和遮挡目标检测效果较好；部署在NVIDIA TITAN V GPU计算机和NVIDIA Xavier嵌入式计算平台上的检测速度分别达到54.6和26.8 fps，实时性好. | | English Summary: | | A deep learning method detection for target with few samples based on YOLOv3 was proposed to solve the problem of small enemy soldiers’ datasets. Data augmentation was used to improve the robustness of the small-sample target detection model, and improve the network structure by connecting the shallow network feature map to the deep network across layers. k-means clustering was used to obtain anchor boxes suitable for soldier target characteristics, and pre-training was used to improve the convergence speed of model training. The results show that the method in this paper has a success rate (mAP) of 85.6% for target detection of enemy soldiers with small enemy soldiers’ datasets, a detection accuracy (IOU) of 82.18%, and a good detection effect for small and occluded targets. The detection speed deployed on NVIDIA TITAN V GPU computer and NVIDIA Xavier reaches 54.6 and 26.8 fps, which means a good real-time performance. | | 查看全文查看/发表评论下载PDF阅读器 | |
王放,邢冀川.一种改进型神经网络的光纤预警系统适应性研究[J].北京理工大学学报（自然科学版）,2021,41(6):649~657.WANGFang,XINGJichuan.AnImprovedNeuralNetworkBasedResearchonGeneralizationofanOptica ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21陈军,兀亚伟,李垣志,钱新明,袁梦琦.基于动态贝叶斯网络的燃气管网燃爆风险分析[J].北京理工大学学报（自然科学版）,2021,41(7):696~705.CHENJun,WUYawei,LIYuanzhi,QIANXinming,YUANMengqi.RiskAnalysisofBurningan ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21张宏伟,达新宇,胡航,倪磊,潘钰.基于协作频谱感知的多无人机通信网络谱效优化研究[J].北京理工大学学报（自然科学版）,2021,41(8):830~839.ZHANGHongwei,DAXinyu,HUHang,NILei,PANYu.SpectrumEfficiencyOptimizationo ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21罗森林,杨俊楠,潘丽敏,吴舟婷.面向信息与通信技术供应链网络画像构建的文本语义匹配方法[J].北京理工大学学报（自然科学版）,2021,41(8):864~872.LUOSenlin,YANGJunnan,PANLimin,WUZhouting.TextSemanticMatchingMethodf ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21郑戍华,南若愚,李守翔,王向周,陈梦心.基于轻量化网络的眼部特征分割方法[J].北京理工大学学报（自然科学版）,2021,41(9):970~976.ZHENGShuhua,NANRuoyu,LIShouxiang,WANGXiangzhou,CHENMengxin.ALightweight-Net ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21刘国满,聂旭娜.一种基于卷积神经网络的雷达干扰识别算法[J].北京理工大学学报（自然科学版）,2021,41(9):990~998.LIUGuoman,NIEXuna.ARadarJammingRecognitionAlgorithmBasedonConvolutionalNeuralNetwork ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21韩子硕,王春平,付强.基于深层次特征增强网络的SAR图像舰船检测[J].北京理工大学学报（自然科学版）,2021,41(9):1006~1014.HANZishuo,WANGChunping,FUQiang.ShipDetectioninSARImagesBasedonDeepFeatureEnha ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21王建中,徐浩楠,王洪枫,于子博.基于残差密集块和自编码网络的红外与可见光图像融合[J].北京理工大学学报（自然科学版）,2021,41(10):1077~1083.WANGJianzhong,XUHaonan,WANGHongfeng,YUZibo.InfraredandVisibleImageFu ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21陈宇,张勇,陈实.大规模卫星集群网络自适应加权分簇算法[J].北京理工大学学报（自然科学版）,2021,41(11):1188~1192.CHENYu,ZHANGYong,CHENShi.AdaptiveWeightedClusteringAlgorithmforLarge-ScaleSatelli ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21王剑,王新民,谢蓉,李婷,曹宇燕.基于IMM-UKF方法的机电作动器突发性故障诊断研究[J].北京理工大学学报（自然科学版）,2019,39(2):198~202,208.WANGJian,WANGXin-ming,XIERong,LITing,CAOYu-yan.AbruptFaultDiagno ... 北京理工大学科研学术 本站小编 Free考研考试 2021-12-21
| |
