|本期目录/Table of Contents|

[1]季 航,贾 镕,刘 晓,等.一种基于YOLOv3的红外目标检测系统[J].电子设计工程,2019,27(22):61-64.
 JI Hang,JIA Rong,LIU Xiao,et al.An infrared target detection system based on YOLOv3[J].SAMSON,2019,27(22):61-64.
点击复制

一种基于YOLOv3的红外目标检测系统(PDF)
分享到:

《电子设计工程》[ISSN:1674-6236/CN:61-1477/TN]

卷:
27
期数:
2019年22期
页码:
61-64
栏目:
计算机技术应用
出版日期:
2019-11-20

文章信息/Info

Title:
An infrared target detection system based on YOLOv3
文章编号:
1674-6236(2019)22-0061-04
作者:
季 航1贾 镕12刘 晓12拓浩男1
(1中国人民解放军陆军炮兵防空兵学院 安徽 合肥 230031;2偏振光成像探测技术安徽省重点实验室安徽 合肥 230031)
Author(s):
JI Hang1JIA Rong12LIU Xiao12TUO Hao?nan1
(1. Army Artillery and Air Defense Forces College, Hefei 230031,China;2. Anhui province Key Laboratory of Polarization Imaging Detection Technology,Hefei 230031,China)
关键词:
深度学习 YOLOv3 红外图像 检测系统 军事目标
Keywords:
deep learning YOLOv3 infrared image detection system military target
分类号:
TP399
DOI:
-
文献标志码:
A
摘要:
针对未来战场感知体系的自动化和智能化的需求,设计出了一种基于深度学习的战场红外目标检测系统。近年来随着深度卷积神经网络在图像识别领域的广泛应用,因此将这项新技术应用于军事目标检测具有极强的现实意义。系统通过红外成像机芯采集红外图像,使用图像采集卡实时传输图像数据,最后在主机端利用深度卷积神经网络进行目标检测。以YOLOv3算法作为基础,以某款金属车辆模型为例,采集该型车辆的红外图像数据并制作相应的数据集,训练得到相应的检测内核。通过实验证明,检测速度可达在30 帧/秒以上,且在fps达到30的前提下平均识别精度达到70%以上,不仅具有很好的实时性、准确性,对各种环境变化具有较好的鲁棒性,验证了该系统的可行性和军事应用价值。
Abstract:
Aiming at the automation and intellectualization of battlefield perception system in the future, a battlefield infrared target detection system based on deep learning is designed in this paper. In recent years, with the extensive application of deep convolution neural network in the field of image recognition, it is of great practical significance to apply this new technology to military target detection. The system collects infrared image through infrared imaging core, transmits image data in real time by image acquisition card, and finally detects target by depth convolution neural network at host end. Based on YOLO V3 network and taking a metal vehicle model as an example, this paper collects the infrared image data of the vehicle and produces the corresponding data sets, and trains to get the corresponding detection core. Experiments show that the detection speed can reach more than 30 frames/seconds, and the average recognition accuracy can reach more than 70% under the premise of FPS reaching 30. It not only has good real-time and accuracy, but also has good robustness to various environmental changes, which verifies the feasibility and military application value of the system.

参考文献/References:

[1] 冯小雨,梅卫,胡大帅. 基于改进Faster R-CNN的空中目标检测[J]. 光学学报, 2018,38(6):250-258.[2] 朱晨阳,冯虎田,欧屹. 基于YOLO3的人脸自动跟踪摄像机器人系统研究[J]. 电视技术, 2018, 42(9):64-69,91.[3] 郑胤,陈权崎,章毓晋. 深度学习及其在目标和行为识别中的新进展[J]. 中国图象图形学报, 2014,19(2):175-184.[4] 朱超平,杨艺. 基于YOLO2和ResNet算法的监控视频中的人脸检测与识别[J]. 重庆理工大学学报(自然科学), 2018,32(8):176-181.[5] 戴伟聪,金龙旭,李国宁,等. 遥感图像中飞机的改进YOLOv3实时检测算法[J]. 光电工程, 2018,45(12):84-92.[6] 张家晨, 陈庆奎. 基于YOLO的道路车辆拥堵分析模型[J]. 计算机应用, 2019,39(1):93-97.[7] Li J, Zhang D, Zhang J, et al. Facial expression recognition with faster R-CNN[J]. Procedia Computer Science, 2017,107(C):135-140.[8] 张楚楚,吕学斌. 基于改进YOLOv2网络的密集人群场景行人检测[J]. 现代计算机(专业版), 2018,628(28):36-41.[9] 黎洲,黄妙华. 基于YOLO_v2模型的车辆实时检测[J]. 中国机械工程, 2018,29(15):109-114.[10]李云鹏,侯凌燕,王超. 基于YOLOv2的复杂场景下车辆目标检测[J]. 电视技术, 2018,42(5):105-111.[11]Sang J, Guo P, Xiang Z, et al. Vehicle detection based on faster-RCNN[J]. Journal of Chongqing University, 2017, 40(7):32-36.[12]熊俊涛,刘振,林睿,等. 自然环境下树上绿色芒果的无人机视觉检测技术[J]. 农业机械学报, 2018,49(11):30-36.[13]Lecun Y, Bengio Y, Hinton G. Deep learning.[J]. Nature, 2015, 521(7553):436.[14]Deng L,Yu D. Deep learning: methods and applications[J]. Foundations & Trends in Signal Processing, 2014,7(3):197-387.[15]强勇,缑水平,王永刚. 战场感知系统目标识别技术的进展[J]. 火控雷达技术, 2008, 37(1):1-9.[16]李美满. 基于红外图像的船舶特征识别方法[J]. 舰船科学技术, 2018,40(12):77-79.

相似文献/References:

[1]花兴艳,葛耀林.基于弱小红外目标的图像增强研究[J].电子设计工程,2016,(24):148.
 HUA Xing-yan,GE Yao-lin.Image enhancement algorithm for the dim targets in infrared images[J].SAMSON,2016,(22):148.
[2]黄荣辉,张宏钊,姚森敬,等.光子型紫外检测管理系统在高压电气设备的应用[J].电子设计工程,2016,(12):146.
 HUANG Rong-hui,ZHANG Hong-zhao,YAO Sen-jing,et al.Photon type ultraviolet detection management system in the application of high voltage electrical equipment[J].SAMSON,2016,(22):146.
[3]陈 茹.电子线路板卡自动检测系统输入输出扩展电路的设计[J].电子设计工程,2016,(12):121.
 CHEN Ru.Electronic circuit board automatically detects the system input and output expansion circuit design[J].SAMSON,2016,(22):121.
[4]李倩云,夏 斌.基于EEG的睡眠数据的分类[J].电子设计工程,2016,(05):26.
 LI Qian-yun,XIA Bin.Sleep classification based on EEG[J].SAMSON,2016,(22):26.
[5]单外平,曾雪琼.基于深度信念网络的信号重构与轴承故障识别[J].电子设计工程,2016,(04):67.
 SHAN Wai-ping,ZENG Xue-qiong.Signal reconstruction and bearing fault identification based on deep belief network[J].SAMSON,2016,(22):67.
[6]钟修皓,王晓虹,刘文壮,等.智能语音巡检机器人系统的设计与实现[J].电子设计工程,2015,(15):5.
 ZHONG Xiu-hao,WANG Xiao-hong,LIU Wen-zhuang,et al.Design and implementation of intelligent voice touring inspection robot system[J].SAMSON,2015,(22):5.
[7]井金剑,翟春艳,李书臣,等.基于红外图像识别的电气设备故障诊断[J].电子设计工程,2014,(12):171.
 JING Jin-jian,ZHAI Chun-yan,LI Shu-chen,et al.Fault diagnosis of electrical equipment based on infrared image recognition[J].SAMSON,2014,(22):171.
[8]李 欣. 基于形态学与二维Otsu相结合的红外图像分割方法[J].电子设计工程,2013,(17):138.
 LI Xin. Infrared image segmentation based on the combination of morphological and two-dimensional Otsu method[J].SAMSON,2013,(22):138.
[9]尉广军,何 鹏,姚志敏,等.基于PC104的某型导弹随动装置检测系统设计[J].电子设计工程,2011,(13):85.
 YU Guang-jun,HE Peng,YAO Zhi-min,et al.Design of missile follow-up mechanism test system based on PC104[J].SAMSON,2011,(22):85.
[10]伟利国,张小超,李福超,等.基于微波的粮食水分检测系统设计[J].电子设计工程,2010,(04):1.
 Design of grain moisture content measurement system based on microwave.The Design of Grain Moisture Content Measurement system based on Microwave[J].SAMSON,2010,(22):1.

备注/Memo

备注/Memo:
收稿日期:2019-03-20 稿件编号:201903146基金项目:国家自然科学基金(41406109);安徽省自然科学基金(1708085QD90)作者简介:季 航(1995—),男,浙江绍兴人,硕士研究生。研究方向:多源信息融合。
更新日期/Last Update: 2019-11-22