Few-shot warhead fragment group object detection based on feature reassembly and attention

HE Meng; WU Jiangpeng; LIANG Chao; HU Pengyu; REN Yuan; HE Xuan; LIU Qianghui

doi:10.37188/OPE.20243212.1929

您当前的位置：

首页 >

文章列表页 >

Few-shot warhead fragment group object detection based on feature reassembly and attention

Information Sciences | 更新时间：2024-07-26

- Few-shot warhead fragment group object detection based on feature reassembly and attention
- Optics and Precision Engineering Vol. 32, Issue 12, Pages: 1929-1940(2024)
- 作者机构：
  
  西安现代控制技术研究所，陕西西安 710065
- 作者简介：
- 基金信息：
- DOI：10.37188/OPE.20243212.1929
  CLC： TP391.4
- Published：25 June 2024，
  
  Received：09 March 2024，
  
  Revised：19 April 2024，
- 稿件说明：
移动端阅览
和萌,武江鹏,梁超等.融合特征重组和注意力的小样本战斗部破片群目标检测[J].光学精密工程,2024,32(12):1929-1940.

HE Meng,WU Jiangpeng,LIANG Chao,et al.Few-shot warhead fragment group object detection based on feature reassembly and attention[J].Optics and Precision Engineering,2024,32(12):1929-1940.
和萌,武江鹏,梁超等.融合特征重组和注意力的小样本战斗部破片群目标检测[J].光学精密工程,2024,32(12):1929-1940. DOI： 10.37188/OPE.20243212.1929.

HE Meng,WU Jiangpeng,LIANG Chao,et al.Few-shot warhead fragment group object detection based on feature reassembly and attention[J].Optics and Precision Engineering,2024,32(12):1929-1940. DOI： 10.37188/OPE.20243212.1929.

摘要

战斗部破片群运动参数对弹药毁伤威力评估具有重要的意义。针对破片尺寸较小、背景信息复杂以及破片数据样本少导致的破片检测精度较低的问题，本文提出一种YOLOv5-FD的战斗部破片群目标检测方法。首先，在网络输出端增加微小目标检测层，将原始的三尺度改为四尺度，并在特征融合网络中引入内容感知特征重组（Content Aware ReAssembly of Features，CARAFE）上采样模块替换原有的最近邻插值上采样，减少小目标特征信息损失，提高弱小破片的提取能力。其次，在特征提取网络引入坐标注意力模块（Coordinate Attention，CA），加强对破片特征的提取，弱化背景信息，抑制复杂背景的干扰。最后，在模型训练过程中引入模型不可知元学习方法（Model Agnostic Meta Learning，MAML），达到仅用小样本破片数据集实现较高的检测性能。实验结果表明，YOLOv5-FD破片检测算法在自制破片数据集中，精确率达到了90.5%，召回率达到了85.4%，平均精度mAP_0.5达到了88.2%，与原始YOLOv5s算法相比分别提高了7.1%，7.9%和7.5%，有效提高了破片目标检测准确性。

Abstract

The motion parameters of warhead fragment group has important significance for evaluating the damage power of ammunition. Aiming at the problems of low fragment object detection precision caused by small fragment object size， complex background information， and few fragment samples， this paper proposed a YOLOv5-FD（You Only Look Once v5-Fragment Detection） method for warhead fragment group object detection. Firstly， a small object detection layer was added at the network output layer， changing the original three-scales to four-scales， and a lightweight upsampling module called CARAFE （Content Aware ReAssembly of FEatures） was introduced in the feature fusion network to replace the original nearest neighbor interpolation upsampling， reducing the loss of small object feature information and improving the ability to extract small fragments. Secondly， the CA （Coordinate Attention） module was introduced into the feature extraction network to enhance of fragment features， weaken background information， and suppress interference from complex backgrounds. Finally， the MAML （Model Agnostic Meta Learning） algorithm was introduced during the model training process to achieve high detection performance using only few-shot fragment datasets. The experimental results show that YOLOv5-FD algorithm achieves the precision of 90.5%， the recall rate of 85.4%， and the average precision of 88.2% in the self-made fragment datasets. Compared with the original YOLOv5s algorithm， it improved by 7.1%， 7.9%， and 7.5%， respectively， effectively improving the precision of fragment object detection.

关键词

高速破片目标检测YOLOv5特征重组注意力机制元学习

Keywords

high-speed fragmentobject detectionYOLOv5feature reassemblyattention mechanismmeta-learning

references

武江鹏，乔明军，闫振纲，等. 战斗部破片场参数测试技术发展综述［J］. 兵器装备工程学报， 2019， 40（5）： 105-109. doi: 10.11809/bqzbgcxb2019.05.023http://dx.doi.org/10.11809/bqzbgcxb2019.05.023

WU J P， QIAO M J， YAN Z G， et al. Overview of testing technology for warhead fragments characterization［J］. Journal of Ordnance Equipment Engineering， 2019， 40（5）： 105-109.（in Chinese）. doi: 10.11809/bqzbgcxb2019.05.023http://dx.doi.org/10.11809/bqzbgcxb2019.05.023

SUN C， JIA Y Y， WANG D N. Modeling of high-speed laser photography system for field projectile testing［J］. Optik， 2021， 241： 166980. doi: 10.1016/j.ijleo.2021.166980http://dx.doi.org/10.1016/j.ijleo.2021.166980

WANG H， XIAO N F. Underwater object detection method based on improved faster RCNN［J］. Applied Sciences， 2023， 13（4）： 2746. doi: 10.3390/app13042746http://dx.doi.org/10.3390/app13042746

李想，特日根，仪锋，等. 针对全球储油罐检测的TCS-YOLO模型［J］. 光学精密工程， 2023， 31（2）： 246-262. doi: 10.37188/OPE.20233102.0246http://dx.doi.org/10.37188/OPE.20233102.0246

LI X， TE R G， YI F， et al. TCS-YOLO model for global oil storage tank inspection［J］. Opt. Precision Eng.， 2023， 31（2）： 246-262.（in Chinese）. doi: 10.37188/OPE.20233102.0246http://dx.doi.org/10.37188/OPE.20233102.0246

ZHU X K， LYU S C， WANG X， et al. TPH-YOLOv5： improved Yolov5 based on transformer prediction head for object detection on drone-captured scenarios［C］. 2021 IEEE/CVF International Conference on Computer Vision Workshops （ICCVW）. Montreal， BC， Canada. IEEE， 2021： 2778-2788. doi: 10.1109/iccvw54120.2021.00312http://dx.doi.org/10.1109/iccvw54120.2021.00312

PENG C， ZHU M， REN H E， et al. Small object detection method based on weighted feature fusion and CSMA attention module［J］. Electronics， 2022， 11（16）： 2546. doi: 10.3390/electronics11162546http://dx.doi.org/10.3390/electronics11162546

杨晨，佘璐，杨璐，等. 改进YOLOv5的遥感影像目标检测算法［J］. 计算机工程与应用， 2023， 59（15）： 76-86. doi: 10.3778/j.issn.1002-8331.2301-0220http://dx.doi.org/10.3778/j.issn.1002-8331.2301-0220

YANG C， SHE L， YANG L， et al. Improved YOLOv5 object detection algorithm for remote sensing images［J］. Computer Engineering and Applications， 2023， 59（15）： 76-86.（in Chinese）. doi: 10.3778/j.issn.1002-8331.2301-0220http://dx.doi.org/10.3778/j.issn.1002-8331.2301-0220

LIAO H， ZHU W Q. YOLO-DRS： a bioinspired object detection algorithm for remote sensing images incorporating a multi-scale efficient lightweight attention mechanism［J］. Biomimetics， 2023， 8（6）： 458. doi: 10.3390/biomimetics8060458http://dx.doi.org/10.3390/biomimetics8060458

LIU L Y， WANG B C， KUANG Z H， et al. GenDet： meta learning to generate detectors from few shots［J］. IEEE Transactions on Neural Networks and Learning Systems， 2022， 33（8）： 3448-3460. doi: 10.1109/tnnls.2021.3053005http://dx.doi.org/10.1109/tnnls.2021.3053005

陈筱，朱向冰，吴昌凡，等. 基于迁移学习与特征融合的眼底图像分类［J］. 光学精密工程， 2021， 29（2）： 388-399. doi: 10.37188/OPE.20212902.0388http://dx.doi.org/10.37188/OPE.20212902.0388

CHEN X， ZHU X B， WU C F， et al. Research on fundus image classification based on transfer learning and feature fusion［J］. Opt. Precision Eng.， 2021， 29（2）： 388-399.（in Chinese）. doi: 10.37188/OPE.20212902.0388http://dx.doi.org/10.37188/OPE.20212902.0388

郭保青，张德芬. 基于度量元学习的铁路小样本入侵目标检测方法［J］. 光学精密工程， 2023， 31（12）： 1816-1826. doi: 10.37188/ope.20233112.1816http://dx.doi.org/10.37188/ope.20233112.1816

GUO B Q， ZHANG D F. Railway few-shot intruding objects detection method with metric meta learning［J］. Opt. Precision Eng.， 2023， 31（12）： 1816-1826.（in Chinese）. doi: 10.37188/ope.20233112.1816http://dx.doi.org/10.37188/ope.20233112.1816

WU J X， TIAN X T， ZHONG G Q. Supervised contrastive representation embedding based on transformer for few-shot classification［J］. Journal of Physics： Conference Series， 2022， 2278（1）： 012022. doi: 10.1088/1742-6596/2278/1/012022http://dx.doi.org/10.1088/1742-6596/2278/1/012022

YAO X， ZHU J L， HUO G Y， et al. Model-agnostic multi-stage loss optimization meta learning［J］. International Journal of Machine Learning and Cybernetics， 2021， 12（8）： 2349-2363. doi: 10.1007/s13042-021-01316-6http://dx.doi.org/10.1007/s13042-021-01316-6

RIEMER M， CASES I， AJEMIAN R， et al. Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference［EB/OL］. 2018： arXiv： 1810.11910. http：//arxiv.org/abs/1810.11910http://arxiv.org/abs/1810.11910

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Aircraft target detection in SAR images based on MA-DETR

Position-sensitive Transformer aerial image object detection model

Railway few-shot intruding objects detection method with metric meta learning

Lightweight small object detection network with attention mechanism

Integrated 2D-3D LiDAR-vision fusion vehicle speed estimation based on image frustum

Related Author

ZHOU Wenjun

HUANG Shuo

ZHANG Ning

SONG Chuanlong

ZHAO Yuxuan

DUAN Yifan

XU Guoqing

LI Daxiang

Related Institution

School of Communication and Information Engineering， Shanghai University

Shanghai Key Laboratory of Space-based Heterogeneous Network Collaborative Computing， Shanghai Aerospace Electronic Communication Equipment Research Institute

College of communication and information engineering， Xi'an University of Posts and Telecommunication

School of Mechanical， Electronic and Control Engineering， Beijing Jiaotong University

Frontiers Science Center for Smart High-Speed Railway System，Beijing Jiaotong University

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰