[an error occurred while processing this directive] [an error occurred while processing this directive] [an error occurred while processing this directive]
[an error occurred while processing this directive]

弹箭与制导学报 >

2021 , Vol. 41 >Issue 6: 53 - 58

DOI: https://doi.org/10.15892/j.cnki.djzdxb.2021.06.011

一种基于YOLOv4改进的军事目标检测方法

  • 郭昊昌 ,
  • 于力 ,
  • 刘镇涛
展开
  • 西安现代控制技术研究所,西安 710065

郭昊昌(1994—),男,安徽六安人,助理工程师,硕士,研究方向:目标识别。

收稿日期: 2020-09-15

  网络出版日期: 2025-02-07

收起

An Object Detection Method for Military Target Based on Improved YOLOv4

  • GUO Haochang ,
  • YU Li ,
  • LIU Zhentao
Expand
  • Xi'an Modern Control Technology Research Institute,Xi'an 710065, China

Received date: 2020-09-15

  Online published: 2025-02-07

Fold

摘要

针对传统的目标检测算法目标背景过于复杂、目标尺度变化过大、目标遮挡预计运动模糊等问题,使用深度学习的目标检测算法来改善。根据自建的军事目标数据集的特点,对最新YOLOv4算法进行网络结构改进,重新计算先验框的数量以及引入空间注意力机制。改进后的网络与其他主流目标检测算法相比达到了较高的 P ¯ m值,同时识别速度F在38.2 帧/s,满足军事目标检测的实时性需求。

关键词： 军事目标检测; 深度学习目标检测算法; YOLOv4; 军事目标数据集

本文引用格式

郭昊昌 , 于力 , 刘镇涛 . 一种基于YOLOv4改进的军事目标检测方法[J]. 弹箭与制导学报, 2021 , 41(6) : 53 -58 . DOI: 10.15892/j.cnki.djzdxb.2021.06.011

Abstract

In order to solve the problem of traditional object detection such as complex background of object; various object scale;object occlusion and motion blur, we use deep learning algorithm for object detection. According to the data set established by the author, the YOLOv4 algorithm is refined in some aspects such as improving the network architecture, changing the number of anchor boxes and using spatial attention module. As the results, the improved YOLOv4 achieves the highest value of mAP comparing with other methods and the frame rate is 38.2 fps which satisfies the requirements of military object detection.

Key words： military object detection; deep learning algorithm; YOLOv4; military target data set

[an error occurred while processing this directive]

参考文献

原文顺序 | 文献年度倒序 | 文中引用次数倒序
[1]
杨传栋, 刘桢, 石胜斌, 等. 基于CNN的弹载图像目标检测方法研究[J]. 战术导弹技术, 2019, 4(4):85-92.

[2]
ZHAO Z Q, ZHENG P X, SHOU T, et al. Object detection with deep learning: a review[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(11): 3212-3232.

[3]
BOCHKOVSKIY A, WANG C Y, LIAO H Y. YOLOv4: optimal speed and accuracy of object detection[D]. Ithaca: Cornell University, 2020.

[4]
WANG C Y, LIAO H Y M, WU Y H, et al. CSPNet: a new backbone that can enhance learning capability of CNN[C]// CVPR. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop. New York: IEEE, 2020: 7961-7972.

[5]
CHANDRA D, SHILPI B, MATANGINI C, et al. A novel distance based modified k-means clustering algorithm for estimation of missing values in micro-array gene expression data[J]. International Journal of Information Technology & Management Information System, 2014, 5(3): 1-13.

[6]
LIU S, QI L, QIN H F, et al. Path aggregation network for instance segmentation[C]// CVPR. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2018: 8759-8768.

[7]
REDMON J, DIVVALA S, GIRSHICK R, et al. A you only look once: unified, real-time object detection[C]// CVPR. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2016: 4721-4730.

[8]
LIN T Y, DOLLAR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]// IEEE. Proceedings of the IEEE conference on computer vision and pattern recognition. New York: IEEE, 2017: 2117-2125.

[9]
HE K M, ZHANG X Y, REN S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1904-1916.

DOI PMID

[10]
RENMON J, FARHADI A. YOLO9000: better, faster, stronger[C]// CVPR. Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2017: 4515-4519.

Options
文章导航

/

〈 〉

[an error occurred while processing this directive]