基于改进YOLOv5的LPI雷达信号检测识别一体化方法

doi:10.12305/j.issn.1001-506X.2026.09.14

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (9): 3008-3018.doi: 10.12305/j.issn.1001-506X.2026.09.14

• 传感器与信号处理 • 上一篇

基于改进YOLOv5的LPI雷达信号检测识别一体化方法

雷蕾¹^,²(), 朱应申¹, 阮航¹, 周东平¹, 付亦凡¹, 潘黎¹, 郭世盛²()

1. 北京无线电测量研究所，北京 100854
2. 电子科技大学信息与通信工程学院，四川成都 611731

收稿日期:2025-05-29 修回日期:2025-08-29 接受日期:2025-11-28 出版日期:2026-02-10 发布日期:2026-02-10
通讯作者: 郭世盛 E-mail:leilei202309@163.com;ssguo@uestc.edu.cn
作者简介:雷　蕾（2000—），女，硕士研究生，主要研究方向为目标探测
朱应申（1985—），男，研究员，硕士，主要研究方向为目标探测
阮　航（1988—），男，高级工程师，博士，主要研究方向为目标探测
周东平（1997—），男，博士研究生，主要研究方向为目标探测
付亦凡（1999—），男，硕士研究生，主要研究方向为目标探测
潘　黎（2001—），男，硕士研究生，主要研究方向为目标探测

Integrated detection and recognition method for LPI radar signals based on improved YOLOv5

Lei Lei¹^,²(), Yingshen Zhu¹, Hang Ruan¹, Dongping Zhou¹, Yifan Fu¹, Li Pan¹, Shisheng Guo²()

1. Beijing Institute of Radio Measurement，Beijing 100854，China
2. School of Information and Communication Engineering，University of Electronic Science and Technology of China，Chengdu 611731，China

Received:2025-05-29 Revised:2025-08-29 Accepted:2025-11-28 Online:2026-02-10 Published:2026-02-10
Contact: Shisheng Guo E-mail:leilei202309@163.com;ssguo@uestc.edu.cn

摘要/Abstract

摘要：

针对现有研究大多独立解决信号检测或识别问题的现状，提出一种基于改进YOLO（you only look once）v5网络的低截获概率雷达信号检测识别一体化方法，有效解决检测与识别分离导致的信息丢失和效率低下等问题。以YOLOv5为框架，通过引入全局上下文网络、双向特征金字塔结构及Ghost模块，在保证计算效率的同时提升网络准确度。实验结果表明，在参数量压缩19%的同时，当信噪比大于−4 dB时检测识别精度能达到99%以上，证明所提算法能够实现低截获概率雷达信号在低信噪比下的高精度一体化检测和识别。

关键词: 低截获概率雷达信号, 信号检测, 调制识别, 改进YOLOv5

Abstract:

Considering that most existing research solves signal detection or recognition problems seporately an integrated method for low probability of detection （LPI） radar signal detection and recognition based on an improved you only look once （YOLO） v5 network is proposed, which effectively solves the problems of information loss and low efficiency caused by the separation of detection and recognition. This method is based on the YOLOv5 framework and introduces a global context network, bidirectional feature pyramid structure, and Ghost module to improve network accuracy while ensuring computational efficiency. Experimental results demonstrate that the proposed method reduces the number of parameters by 19%, while achieving detection and recognition accuracy exceeding 99% at signal-to-noise ratios above −4 dB which confirms that the proposed algorithm enables high-precision integrated detection and recognition of LPI radar signals under low signal-to-noise conditions.

Key words: low probability of interception (LPI) radar signal, signal detection, modulation recognition, improved you only look once （YOLO） v5

中图分类号:

TN 974

雷蕾, 朱应申, 阮航, 周东平, 付亦凡, 潘黎, 郭世盛. 基于改进YOLOv5的LPI雷达信号检测识别一体化方法[J]. 系统工程与电子技术, 2026, 48(9): 3008-3018.

Lei Lei, Yingshen Zhu, Hang Ruan, Dongping Zhou, Yifan Fu, Li Pan, Shisheng Guo. Integrated detection and recognition method for LPI radar signals based on improved YOLOv5[J]. Systems Engineering and Electronics, 2026, 48(9): 3008-3018.

图/表 15

图1

图2

图3

图4

图5

图6

图7

图8

表1

图9

表2

表3

图10

图11

表4

参考文献 32

1	Ren F T, Quan D Y, Shen L, et al. LPI radar signal recognition based on feature enhancement with deep metric learning[J]. Electronics, 2023, 12 (24): 44934.
2	康智. 基于U-Net的雷达脉冲信号一体化检测、分选和识别[J]. 电子信息对抗技术, 2023, 38 (6): 45. doi: 10.3969/j.issn.1674-2230.2023.06.007
3	滑伟. 基于能量检测的自适应频谱检测算法研究[D].西安: 西安电子科技大学, 2015.
4	Huang H, Yuan C W. A sequential cooperative spectrum sensing algorithm based on dynamic adaptive double-threshold energy detection[J]. Journal of lectronics and Information Technology, 2018, 40 (5): 1037.
5	李渝, 吴涛, 杨文海, 等. 一种分布式目标滑窗检测方法[J]. 空间电子技术, 2022, 19 (2): 64. doi: 10.3969/j.issn.1674-7135.2022.02.009
6	Liu Y J, Xiao P, Wu H C. LPI radar signal detection based on radial integration of Choi-Williams time-frequency image[J]. Journal of Systems Engineering and Electronics, 2015, 26 (5): 973. doi: 10.1109/JSEE.2015.00106
7	王红卫, 范翔宇, 陈游, 等. 基于切片熵权的WHT多LFM信号识别方法[J]. 航空学报, 2017, 38 (1): 214. doi: 10.7527/S1000-6893.2016.0134
8	Chung C D, Polydoros A. Parameter estimation of random FH signals using autocorrelation techniques[J]. IEEE Trans. on Communications, 1995, 43 (2): 1097.
9	Zhang H Y, Guo B, Zhang Y H, et al. Fast cyclostationary feature detection method based on CUDA[J]. Journal of Physics: Conference Series, 2021(1): 1966.
10	Ghadimi G, Norouzi Y, Bayderkhani R, et al. Deep learning-based approach for low probability of intercept radar signal detection and classification[J]. Journal of Communications Technology and Electronics, 2020, 65 (10): 1179. doi: 10.1134/S1064226920100034
11	Hiremath S, Deshmukh S, Rakesh R, et al. Blind identification of radio access techniques based on time-frequency analysis and convolutional neural network[C]//TENCON IEEE Region 10 Conference, 2018: 1163.
12	Prasad V, Dsouza B, BHARGAVA K, et al. A deep learning framework for blind time-frequency localization in wideband systems[C]//IEEE 91st Vehicular Technology Conference, 2020.
13	Ni T, Ding X J, Wang Y F, et al. Spectrum sensing via temporal convolutional network[J]. China Communications, 2021, 18 (9): 37. doi: 10.23919/JCC.2021.09.004
14	Kocaadam E, Ozkazanç Y. Classification of polyphase and polytime LPI radar signals with eigenimage methods[C]//IEEE 18th Signal Processing and Communications Applications Conference, 2010: 173.
15	Xie L J, Wan Q. Cyclic feature based modulation recognition using compressive sensing[J]. IEEE Wireless Communications Letters, 2017, 6 (3): 402. doi: 10.1109/LWC.2017.2697853
16	Jajoo G, Kumar Y, Yadav S K, et al. Blind signal modulation recognition through clustering analysis of constellation signature[J]. Expert Systems with Applications, 2017 (90): 13.
17	李红光, 郭英, 眭萍, 等. 基于时频能量谱纹理特征的跳频调制方式识别[J]. 通信学报, 2019, 40 (10): 20. doi: 10.11959/j.issn.1000-436x.2019191
18	Zhang M, Liu L T, Diao M. LPI radar waveform recognition based on time-frequency distribution[J]. Sensors, 2016, 16 (10): 1682. doi: 10.3390/s16101682
19	Guo Q, Yu X, Ruan G. LPI radar waveform recognition unit based on deep learning transfer learning[J]. Symmetry, 2019, 11, 540. doi: 10.3390/sym11040540
20	Ghadimi G, Norouzi Y, et al. Deep learning-based approach for low probability of intercept radar signal detection and classification[J]. Journal of Communications Technology and Electronics, 2020, 65 (10): 1170.
21	董章华, 赵士杰, 赖莉. 一种基于Swin Transformer神经网络的低截获概率雷达信号调制类型的识别方法[J]. 四川大学学报（自然科学版）, 2023, 60 (2): 42. doi: 10.19907/j.0490-6756.2023.021004
22	Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection[C]//Computer Vision and Pattern Recognition, 2016.
23	Wang X L, Girshick R, Gupta A, et al. Non-local neural networks[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018: 7794.
24	Cao Y, Xu J R, Lin S, et al. GCNet: non-local networks meet squeeze-excitation networks and beyond[C]//IEEE/CVF International Conference on Computer Vision Workshop, 2019: 1971.
25	Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018: 7132.
26	Tan M X, Pang R M, Le Q V. EfficientDet: scalable and efficient object detection[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 10778.
27	Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2017: 936.
28	Han K, Wang Y H, Tian Q, et al. GhostNet: more features from cheap operations[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 1577.
29	Krizheasky A, Sutskever I, Hinton G. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60 (6): 84. doi: 10.1145/3065386
30	Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[C]//3rd International Conference on Learning Representations, 2015.
31	Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2015.
32	He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770.

调制类型	参数设置	数值
LFM NLFM	信号带宽/MHz	30~200
StepF-LFM Costas-LFM	子脉冲带宽/MHz	5~30
	跳频间隔/MHz	10~30
	频率编码个数	{4,5,6,7}
StepF-BPSK Costas-BPSK	跳频间隔/MHz	10~30
	频率编码个数	{4,5,6,7}
	巴克码长度	{7,11,13}
StepF-MPSK Costas-MPSK	跳频间隔/MHz	10~30
	频率编码个数	{4,5,6,7}
	相位控制数	{4,5,6}

参数	信噪比/dB
参数	−10	−8	−6	−4	−2	0	2	4	6
r	57.9	93.5	98.2	99.5	99.6	99.5	99.8	99.6	100.0
p	60.4	86.9	94.9	97.8	99.3	99.3	99.4	99.2	99.2

网络	每秒帧数	浮点计算量（×10⁹）	参数量（×10⁶）
改进的YOLOv5	135.2	41.3	37.2
YOLOv5	125.3	53.9	46.2
AlexNet	1238.6	1.1	58.3
VGG16	751.1	15.4	21.2
GoogleNet	192.3	1.6	5.9
RseNet18	358.9	1.8	11.2

GCNet	BiFPN	GhostNet	r/%	p/%	模型权重/MiB
×	×	×	97.8	97.1	88.5
√	×	×	97.9	97.8	96.6
×	√	×	97.8	97.4	89.0
×	×	√	98.1	97.5	63.1
√	√	×	98.3	98.5	97.1
√	√	√	99.0	98.8	71.7

[1]	胡国梁, 潘继飞, 刘方正, 龚阳, 孙兵, 郭林青. 低信噪比条件下雷达调制样式识别的多任务学习算法[J]. 系统工程与电子技术, 2026, 48(9): 2939-2948.
[2]	王鹏昊, 何迪, 尤明懿, 郭细平. 基于随机共振和多尺度特征融合调制识别方法[J]. 系统工程与电子技术, 2026, 48(6): 2096-2104.
[3]	连家威, 张晓林, 颜品, 孙溶辰. 基于小波能量比和改进阈值函数的通信信号降噪算法[J]. 系统工程与电子技术, 2025, 47(7): 2098-2109.
[4]	周东平, 阮航, 沙明辉, 王崇宇, 崔念强, 鲁耀兵. 基于特征融合的Fast-CNN的复杂波形调制识别[J]. 系统工程与电子技术, 2025, 47(7): 2165-2175.
[5]	代振, 韩庆伟, 王平波, 贾文抖. 脉冲噪声中微弱信号稳健检测研究[J]. 系统工程与电子技术, 2025, 47(11): 3551-3558.
[6]	孙鹏, 魏新平, 王恒, 王玮, 巩克现. 基于多特征融合的卫星单-混信号识别方法[J]. 系统工程与电子技术, 2025, 47(11): 3859-3867.
[7]	王海英, 张群英, 成文海, 董家铭, 刘小军. LPI雷达信号调制识别及参数估计研究进展[J]. 系统工程与电子技术, 2024, 46(6): 1908-1924.
[8]	李晓辉, 万宏杰, 石明利, 王先文. 基于特征值的动态数字信道化子带检测算法[J]. 系统工程与电子技术, 2024, 46(5): 1801-1809.
[9]	邹涵, 张天骐, 马焜然, 杨宗方. 基于多特征融合的MIMO-OFDM系统单-混信号调制识别算法[J]. 系统工程与电子技术, 2024, 46(4): 1456-1465.
[10]	平嘉蓉, 李赛, 林云航. Alpha稳定分布噪声和多径干扰下的无人机集群MIMO信号调制识别[J]. 系统工程与电子技术, 2024, 46(11): 3920-3929.
[11]	邓志安, 王治国, 王盛鳌, 司伟建. 同步提取变换去噪的雷达信号调制识别方法[J]. 系统工程与电子技术, 2024, 46(10): 3334-3346.
[12]	陈洋, 廖灿辉, 张锟, 刘建, 王鹏举. 基于自监督对比学习的信号调制识别算法[J]. 系统工程与电子技术, 2023, 45(4): 1200-1206.
[13]	汪锐, 张天骐, 安泽亮, 王雪怡, 方竹. 基于联合特征参数和一维CNN的MIMO-OFDM系统调制识别算法[J]. 系统工程与电子技术, 2023, 45(3): 902-912.
[14]	龚佩, 李天昀, 章昕亮, 寸陈韬. 利用联合特征参数的卫星单-混信号调制识别[J]. 系统工程与电子技术, 2023, 45(2): 589-596.
[15]	翟茹萍, 张书衡, 平嘉蓉. 复杂多径环境下的无人机集群通信波形识别[J]. 系统工程与电子技术, 2023, 45(10): 3312-3320.

基于改进YOLOv5的LPI雷达信号检测识别一体化方法

Integrated detection and recognition method for LPI radar signals based on improved YOLOv5

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 15

参考文献 32

相关文章 15

编辑推荐

Metrics

本文评价