A counter-UAV system based on deep learning and EO detection

Ming-yu YANG; Hao WANG; Han-yu WANG

doi:10.37188/CJLCD.2020-0325

您当前的位置：

首页 >

文章列表页 >

A counter-UAV system based on deep learning and EO detection

Image Processing | 更新时间：2021-09-10

- A counter-UAV system based on deep learning and EO detection
- Chinese Journal of Liquid Crystals and Displays Vol. 36, Issue 9, Pages: 1323-1330(2021)
- 作者机构：
  
  1.中国科学院长春光学精密机械与物理研究所中国科学院航空光学成像与测量重点实验室, 吉林长春 130033
- 作者简介：
- 基金信息：
  
  National Key R&D Program of China(2017YFC0822402)
- DOI：10.37188/CJLCD.2020-0325
  CLC：
扫描看全文
Ming-yu YANG, Hao WANG, Han-yu WANG. A counter-UAV system based on deep learning and EO detection. [J]. Chinese Journal of Liquid Crystals and Displays 36(9):1323-1330(2021)
DOI：

Ming-yu YANG, Hao WANG, Han-yu WANG. A counter-UAV system based on deep learning and EO detection. [J]. Chinese Journal of Liquid Crystals and Displays 36(9):1323-1330(2021) DOI： 10.37188/CJLCD.2020-0325.

摘要

无人机的无序“黑飞”带来一系列安全及社会问题，如何有效地对无人机进行探测、识别甚至打击是当今研究的热点与难点。为此，本文首先搭建了基于转台与高清可见光相机的无人机目标实时光电探测系统，并构建了一个由40 000帧无人机图像组成的样本库；其次，为了更远、更早地发现目标，在YOLOv3模型基础上增加更小的特征尺度，使得模型对小目标检测效果提升；最后，为了实现无人值守、全自动式无人机目标的探测与跟踪，提出一种基于更小特征尺度的YOLOv3与KCF相结合的模型，并通过外场试验确定了无人机目标跟踪过程中，目标丢失时所对应的阈值参数的选取。结果表明，通过在包含大疆御Pro、精灵3等无人机在内的8 000帧无人机图像组成的静态测试集上进行实验，增加更小特征尺度后的模型对小目标的识别率较之原始YOLOv3模型提高约5%；对于1 280×720分辨率的动态视频，每帧检测时间为0.025~0.030 s（33 fps），且根据选定的阈值，当无人机目标丢失后可重新进行检测，每帧跟踪时间为0.010~0.012 s（85 fps），验证了所提方法的有效性，并可满足工程应用中对实时处理的需求。

Abstract

In view of the security and social problems caused by the unauthorized flying of civil UAV, a real-time detection system of UAV target based on a turntable and high-definition visible light camera is built firstly, and a sample database composed of 40 000 UAV images is constructed. Secondly, in order to find targets farther and earlier, a smaller scale of the feature map is added on the basis of YOLOv3 model to improve the detection effect of smaller targets. Finally, a model based on the combination of smaller feature scale for YOLOv3 and KCF model is proposed to detect and track UAV targets automatically, and the selection of the threshold corresponding to target loss in UAV tracking process is determined through field tests. Results demonstrate that the recognition rate of the modified model is 5% higher than that with YOLOv3 model on small targets in the test set composed of 8 000 UAV images including DJI Mavic pro, Phantom 3,etc,. For 1 280×720 resolution video tests, the detection time of each frame is 0.025~0.030 s (33 fps), and according to the selected threshold, when the UAV target is lost, it can be detected automatically, with the tracking time of each frame 0.010~0.012 s (85 fps), which verifies the effectiveness of the proposed method, and meets the needs of real-time processing in engineering applications.

关键词

无人机防范光电探测深度学习YOLOv3模型KCF模型

Keywords

counter-UAVEO detectiondeep learningYOLOv3 modelKCF model

references

蒋兆军 , 成孝刚 , 彭雅琴 , 等 . 基于深度学习的无人机识别算法研究 . 电子技术应用 , 2017 . 43 ( 7 ): 84 - 87 . 10.16157/j.issn.0258-7998.2017.07.022 10.16157/j.issn.0258-7998.2017.07.022 .

Z J JIANG , X G CHENG , Y Q PENG , 等 . A novel UAV recognition algorithm based on deep learning approach . Application of Electronic Technique , 2017 . 43 ( 7 ): 84 - 87 . 10.16157/j.issn.0258-7998.2017.07.022 10.16157/j.issn.0258-7998.2017.07.022 .

张静 , 张科 , 王靖宇 , 等 . 低空反无人机技术现状与发展趋势 . 航空工程进展 , 2018 . 9 ( 1 ):1 - 8, 34 . 10.16615/j.cnki.1674-8190.2018.01.001 10.16615/j.cnki.1674-8190.2018.01.001 .

J ZHANG , K ZHANG , J Y WANG , 等 . A survey on anti-UAV technology and its future trend . Advances in Aeronautical Science and Engineering , 2018 . 9 ( 1 ): 1 - 8, 34 . 10.16615/j.cnki.1674-8190.2018.01.001 10.16615/j.cnki.1674-8190.2018.01.001 .

陈唯实 , 刘佳 , 陈小龙 , 等 . 基于运动模型的低空非合作无人机目标识别 . 北京航空航天大学学报 , 2019 . 45 ( 4 ): 687 - 694 . DOI:10.13700/j.bh.1001-5965.2018.0447http://doi.org/10.13700/j.bh.1001-5965.2018.0447.

W S CHEN , J LIU , X L CHEN , 等 . Non-cooperative UAV target recognition in low-altitude airspace based on motion model . Journal of Beijing University of Aeronautics and Astronautics , 2019 . 45 ( 4 ): 687 - 694 . DOI:10.13700/j.bh.1001-5965.2018.0447http://doi.org/10.13700/j.bh.1001-5965.2018.0447.

王靖宇 , 王霰禹 , 张科 , 等 . 基于深度神经网络的低空弱小无人机目标检测研究 . 西北工业大学学报 , 2018 . 36 ( 2 ): 258 - 263 . DOI:10.3969/j.issn.1000-2758.2018.02.008http://doi.org/10.3969/j.issn.1000-2758.2018.02.008.

J Y WANG , X Y WANG , K ZHANG , 等 . Small UAV target detection model based on deep neural network . Journal of Northwestern Polytechnical University , 2018 . 36 ( 2 ): 258 - 263 . DOI:10.3969/j.issn.1000-2758.2018.02.008http://doi.org/10.3969/j.issn.1000-2758.2018.02.008.

翟进有 , 代冀阳 , 王嘉琦 , 等 . 深度残差网络的无人机多目标识别 . 图学学报 , 2019 . 40 ( 1 ): 158 - 164 . 10.11996/JG.j.2095-302X.2019010158 10.11996/JG.j.2095-302X.2019010158.

J Y ZHAI , J Y DAI , J Q WANG , 等 . Multi-objective identification of UAV based on deep residual network . Journal of Graphics , 2019 . 40 ( 1 ): 158 - 164 . 10.11996/JG.j.2095-302X.2019010158 10.11996/JG.j.2095-302X.2019010158.

虞晓霞 , 刘智 , 耿振野 , 等 . 一种基于深度学习的禁飞区无人机目标识别方法 . 长春理工大学学报(自然科学版) , 2018 . 41 ( 3 ): 95 - 101 . DOI:10.3969/j.issn.1672-9870.2018.03.021http://doi.org/10.3969/j.issn.1672-9870.2018.03.021.

X X YU , Z LIU , Z Y GENG , 等 . A UAV target recognition method for no flying zone based on deep learning . Journal of Changchun University of Science and Technology (Natural Science Edition) , 2018 . 41 ( 3 ): 95 - 101 . DOI:10.3969/j.issn.1672-9870.2018.03.021http://doi.org/10.3969/j.issn.1672-9870.2018.03.021.

祁蒙 , 王林 , 赵柱 , 等 . 新型"低慢小"目标探测处置系统的体系建设 . 激光与红外 , 2019 . 49 ( 10 ): 1155 - 1158 . DOI:10.3969/j.issn.1001-5078.2019.10.001http://doi.org/10.3969/j.issn.1001-5078.2019.10.001.

M QI , L WANG , Z ZHAO , 等 . Advanced low-slow-small targets detection and disposal system construction . Laser & Infrared , 2019 . 49 ( 10 ): 1155 - 1158 . DOI:10.3969/j.issn.1001-5078.2019.10.001http://doi.org/10.3969/j.issn.1001-5078.2019.10.001.

屈旭涛 , 庄东晔 , 谢海斌 . "低慢小"无人机探测方法 . 指挥控制与仿真 , 2020 . 42 ( 2 ): 128 - 135 . DOI:10.3969/j.issn.1673-3819.2020.02.024http://doi.org/10.3969/j.issn.1673-3819.2020.02.024.

X T QU , D Y ZHUANG , H B XIE . Detection methods for low-slow-small (LSS) UAV . Command Control & Simulation , 2020 . 42 ( 2 ): 128 - 135 . DOI:10.3969/j.issn.1673-3819.2020.02.024http://doi.org/10.3969/j.issn.1673-3819.2020.02.024.

陶磊 , 洪韬 , 钞旭 . 基于YOLOv3的无人机识别与定位追踪 . 工程科学学报 , 2020 . 42 ( 4 ): 463 - 468 . DOI:10.13374/j.issn2095-9389.2019.09.10.002http://doi.org/10.13374/j.issn2095-9389.2019.09.10.002.

L TAO , T HONG , X CHAO . Drone identification and location tracking based on YOLOv3 . Chinese Journal of Engineering , 2020 . 42 ( 4 ): 463 - 468 . DOI:10.13374/j.issn2095-9389.2019.09.10.002http://doi.org/10.13374/j.issn2095-9389.2019.09.10.002.

G E HINTON , R R SALAKHUTDINOV . Reducing the dimensionality of data with neural networks . Science , 2006 . 313 ( 5786 ): 504 - 507 . DOI:10.1126/science.1127647http://doi.org/10.1126/science.1127647.

范丽丽 , 赵宏伟 , 赵浩宇 , 等 . 基于深度卷积神经网络的目标检测研究综述 . 光学精密工程 , 2020 . 28 ( 5 ): 1152 - 1164 . DOI:10.3788/OPE.20202805.1152http://doi.org/10.3788/OPE.20202805.1152.

L L FAN , H W ZHAO , H Y ZHAO , 等 . Survey of target detection based on deep convolutional neural networks . Optics and Precision Engineering , 2020 . 28 ( 5 ): 1152 - 1164 . DOI:10.3788/OPE.20202805.1152http://doi.org/10.3788/OPE.20202805.1152.

梁华 , 宋玉龙 , 钱锋 , 等 . 基于深度学习的航空对地小目标检测 . 液晶与显示 , 2018 . 33 ( 9 ): 793 - 800 . DOI:10.3788/YJYXS20183309.0793http://doi.org/10.3788/YJYXS20183309.0793.

H LIANG , Y L SONG , F QIAN , 等 . Detection of small target in aerial photography based on deep learning . Chinese Journal of Liquid Crystals and Displays , 2018 . 33 ( 9 ): 793 - 800 . DOI:10.3788/YJYXS20183309.0793http://doi.org/10.3788/YJYXS20183309.0793.

马浩鹏 , 朱春媚 , 周文辉 , 等 . 基于深度学习的乳液泵缺陷检测算法 . 液晶与显示 , 2019 . 34 ( 1 ): 81 - 89 . DOI:10.3788/YJYXS20193401.0081http://doi.org/10.3788/YJYXS20193401.0081.

H P MA , C M ZHU , W H ZHOU , 等 . Defect detection algorithm of lotion pump based on deep learning . Chinese Journal of Liquid Crystals and Displays , 2019 . 34 ( 1 ): 81 - 89 . DOI:10.3788/YJYXS20193401.0081http://doi.org/10.3788/YJYXS20193401.0081.

REDMON J, FARHADI A. YOLOv3: an incremental improvement[J].arXiv: 1804.02767v1, 2018.

马永杰 , 宋晓凤 . 基于YOLO和嵌入式系统的车流量检测 . 液晶与显示 , 2019 . 34 ( 6 ): 613 - 618 . DOI:10.3788/YJYXS20193406.0613http://doi.org/10.3788/YJYXS20193406.0613.

Y J MA , X F SONG . Vehicle flow detection based on YOLO and embedded system . Chinese Journal of Liquid Crystals and Displays , 2019 . 34 ( 6 ): 613 - 618 . DOI:10.3788/YJYXS20193406.0613http://doi.org/10.3788/YJYXS20193406.0613.

郑欣 , 田博 , 李晶晶 . 基于YOLO模型的宫颈细胞簇团智能识别方法 . 液晶与显示 , 2018 . 33 ( 11 ): 965 - 971 . DOI:10.3788/YJYXS20183311.0965http://doi.org/10.3788/YJYXS20183311.0965.

X ZHENG , B TIAN , J J LI . Intelligent recognition method of cervical cell cluster based on YOLO model . Chinese Journal of Liquid Crystals and Displays , 2018 . 33 ( 11 ): 965 - 971 . DOI:10.3788/YJYXS20183311.0965http://doi.org/10.3788/YJYXS20183311.0965.

方明 , 孙腾腾 , 邵桢 . 基于改进YOLOv2的快速安全帽佩戴情况检测 . 光学精密工程 , 2019 . 27 ( 5 ): 1196 - 1205 . DOI:10.3788/OPE.20192705.1196http://doi.org/10.3788/OPE.20192705.1196.

M FANG , T T SUN , Z SHAO . Fast helmet-wearing-condition detection based on improved YOLOv2 . Optics and Precision Engineering , 2019 . 27 ( 5 ): 1196 - 1205 . DOI:10.3788/OPE.20192705.1196http://doi.org/10.3788/OPE.20192705.1196.

巫明秀 , 吴谨 , 张晨 , 等 . 基于改进YOLOv3的棉花异性纤维检测 . 液晶与显示 , 2020 . 35 ( 11 ): 1195 - 1203 . DOI:10.37188/YJYXS20203511.1195http://doi.org/10.37188/YJYXS20203511.1195.

M X WU , J WU , C ZHANG , 等 . Detection of foreign fiber in cotton based on improved YOLOv3 . Chinese Journal of Liquid Crystals and Displays , 2020 . 35 ( 11 ): 1195 - 1203 . DOI:10.37188/YJYXS20203511.1195http://doi.org/10.37188/YJYXS20203511.1195.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Infrared dim-small target detection under complex background based on attention mechanism

FRKDNet： feature refine semantic segmentation network based on knowledge distillation

Image rain removal algorithm based on multi-cascade progressive convolution structure

Optical flow estimation via fusing sequence image intensity correlation information

Related Author

No data

Related Institution

Key Laboratory of Airborne Optical Imaging and Measurement， Changchun Institute of Optics，Fine Mechanics and Physics， Chinese Academy of Science，， Changchun

Changchun Institute of Optics， Fine Mechanics and Physics， Chinese Academy of Sciences

University of Chinese Academy of Sciences

College of Big Data and Information Engineering， Guizhou University

Guiyang Aluminum-magnesium Design and Research Institute Co.Ltd.

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
Tel：0431-86176059 Email：yjxs@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