基于RSPWVD重排谱图算法的无人机识别研究

doi:10.12305/j.issn.1001-506X.2026.11.15

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (11): 3836-3849.doi: 10.12305/j.issn.1001-506X.2026.11.15

基于RSPWVD重排谱图算法的无人机识别研究

葛立豪(), 李坤(), 刘帅杰, 潘孟冠, 潘雪莉, 杨利霞, 廖桂生

安徽大学电子信息工程学院，安徽合肥 230601

收稿日期:2025-10-11 修回日期:2025-12-23 接受日期:2026-02-28 出版日期:2026-10-25 发布日期:2026-04-27
通讯作者: 李坤 E-mail:1981283287@qq.com;lik@ahu.edu.cn
作者简介:葛立豪（2001—），男，硕士研究生，主要研究方向为雷达信号处理
刘帅杰（2001—），男，硕士研究生，主要研究方向为雷达信号处理
潘孟冠（1992—），男，讲师，博士，主要研究方向为无线定位、雷达信号处理
潘雪莉（1990—），女，副教授，博士，主要研究方向为雷达信号处理
杨利霞（1975—），男，教授，博士，主要研究方向为雷达信号处理
廖桂生（1963—），男，教授，博士，主要研究方向为雷达信号处理
基金资助:
安徽省重点研发计划项目（2023Z04020018）；安徽省高校优秀科研创新团队（2024AH010003）；国家自然科学基金（62201004）资助课题

UAV recognition research using RSPWVD spectrogram reassignment algorithm

Lihao Ge(), Kun Li(), Shuaijie Liu, Mengguan Pan, Xueli Pan, Lixia Yang, Guisheng Liao

School of Electronic and Information Engineering，Anhui University，Hefei 230601，China

Received:2025-10-11 Revised:2025-12-23 Accepted:2026-02-28 Online:2026-10-25 Published:2026-04-27
Contact: Kun Li E-mail:1981283287@qq.com;lik@ahu.edu.cn

摘要/Abstract

摘要：

针对旋翼无人机识别中提取微多普勒特征问题，提出重排平滑伪Wigner-Ville分布方法。在传统Wigner-Ville分布基础上引入平滑核函数抑制交叉项干扰，并结合谱图重排技术实现能量锐化，从而在抑制交叉项的同时保持良好的时频分辨率与能量聚集性，并分散噪声以突出信号。实验表明，重排平滑伪Wigner-Ville分布在能量聚集和时频分辨率上优于传统方法。将生成的时频图输入YOLOv5模型进行分类，实现了99.7%的识别准确率，证明所提方法在无人机识别中具有有效性与先进性。

关键词: 微多普勒特征, 谱图重排, 重排平滑伪Wigner-Ville分布, YOLOv5

Abstract:

For extrating micro-Doppler signatures in identifying rotor-based unmanned aerial vehicles （UAVs）, a reassigned smoothed pseudo Wigner-Ville distribution （RSPWVD） method is proposed. Building upon the conventional Wigner-Ville distribution, a smoothing kernel is introduced to suppress cross-term interference while incorporating spectrogram reassignment to achieve energy sharpening, significantly suppressing cross-terms while maintaining high time-frequency resolution and energy concentration, and the noise is dispersed to make signal components more distinct. Experimental results demonstrate that RSPWVD outperforms traditional time-frequency analysis methods in both energy concentration and time-frequency resolution. The time-frequency images generated using RSPWVD are fed into a YOLOv5 model for classification, achieving a recognition accuracy of 99.7%, validating the effectiveness and advancement of the proposed approach for UAV recognition.

Key words: micro-Doppler signature, spectrogram reassignment, reassigned smoothed pseudo Wigner-Ville distribution (RSPWVD), YOLOv5

中图分类号:

TN 957

葛立豪, 李坤, 刘帅杰, 潘孟冠, 潘雪莉, 杨利霞, 廖桂生. 基于RSPWVD重排谱图算法的无人机识别研究[J]. 系统工程与电子技术, 2026, 48(11): 3836-3849.

Lihao Ge, Kun Li, Shuaijie Liu, Mengguan Pan, Xueli Pan, Lixia Yang, Guisheng Liao. UAV recognition research using RSPWVD spectrogram reassignment algorithm[J]. Systems Engineering and Electronics, 2026, 48(11): 3836-3849.

图/表 17

图1

图2

图3

图4

表1

图5

图6

表2

表3

图7

表4

图8

表5

图9

表6

图10

表7

参考文献 33

1	王培, 李杨, 崔根, 等. 多旋翼无人机: 新设计、新应用及新发展[J]. 人工智能, 2021 (4): 78. doi: 10.16453/j.cnki.ISSN2096-5036.2021.04.008
2	赵晓琛, 赵东涛, 袁航, 等. 低脉冲重复频率件下无人机微动参数提取[J]. 系统工程与电子技术, 2024, 46 (5): 1503. doi: 10.12305/j.issn.1001-506X.2024.05.05
3	Sornek K, Nadzieja A J, Rosikon I, et al. Status and development prospects of solar-powered unmanned aerial vehicles-a literature review[J]. Energies, 2025, 18 (8): 1924.
4	Han L X, Feng C Q. Micro-Doppler-based space target recognition with a one-dimensional parallel network[J]. International Journal of Antennas and Propagation, 2020, 2020 (1): 1. doi: 10.1155/2020/8013802
5	Schuck T M, Freisel M A, Reese D B. et al. Phenomenology based missile classification using radar micro-Doppler processing [C]//IEEE International Radar Conference, 2020: 192.
6	Chuan C D, Zhao M, Lin J, et al. Transient feature extraction of encoder signal for condition assessment of planetary gearboxes with variable rotational speed[J]. Measurement, 2020, 151, 107206.
7	维婷, 梁青, 谷静. 基于FMCW雷达的旋转目标微动参数提取方法[J]. 系统工程与电子技术, 2021, 43 (12): 3564. doi: 10.12305/j.issn.1001-506X.2021.12.18
8	唐明磊, 张文鹏, 姜卫东, 等. 基于多分辨率显著性滤波的微动信号增强方法[J]. 系统工程与电子技术, 2022, 44 (4): 1148.
9	Tao H, Yi L, Fang C Q, et al. Extracting micro-Doppler features from multi-rotor unmanned aerial vehicles using time-frequency rotation domain concentration[J]. Drones, 2024, 8 (1): 20.
10	Chen V C. Analysis of radar micro-Dopler with time-frequency transform [C]//10th IEEE Workshop on Statistical Signal and Array Processing, 2000: 463.
11	Bin T, Qi Y, Ye Z, et al. Three-dimensional micro-motion feature extraction of the vibrating target based on multi-channel radar in the terahertz band[J]. Sensors, 2019, 20 (1): 8.
12	Max C S, Claire E H, Liam T. Mechanisms of lateral ankle ligament sprainsin professional netball: a systematic video analysis[J]. Physical Therapy in Sport, 2024, 69, 8.
13	Biswas M H A, Massopust P, Ramakrishnan R. The theory of deep convolutional neural networks and a data approximation problem based on the fractional Fourier transform[J]. Applied and Computational Harmonic Analysis, 2025, 81, 101823.
14	Qi M, Su Y, Wang Z, et al. Research on the integration of sensing and communication based on fractional-order Fourier transform[J]. Sensors, 2025, 25 (10): 2956.
15	Liang Q Y, Zhang N H, Wang Q, et al. Identifying and tracking method for the beacon in intersatellite laser communication based on the fractional Fourier transform[J]. Applied Optics, 2025, 64 (26): 7636. doi: 10.1364/AO.561411
16	Liang G L, Guo S X, Zou N, et al. A characteristic extraction method for voice print slice statistics base on joint time-frequency processing[J]. Applied Acoustics, 2024, 216, 109814.
17	Brynolfsson J, Sandsten M. Parameter estimation of oscillating Gaussian functions using the scaled reassigned spectrogram[J]. Signal Processing, 2018 (45): 15020.
18	Chen V C. Analysis of radar micro-Doppler signature with time-frequency transform [C]//10th IEEE Workshop on Statistical Signal and Array Processing, 2000: 463.
19	Chen V C, Li F, Ho S S. et al. Micro-Doppler effect in radar: phenomenon, model and simulation study[J]. IEEE Trans. on Aerospace and Electronic Systems, 2006, 42 (1): 2.
20	Kwan B K, Seong H K, Ook S P. Drone classification using convolutional neural networks with merged Doppler images[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14 (1): 38.
21	Patel J S, Al-Ameri C, Fioranelli F, et al. Multi-time frequency analysis and classificati on of a micro-drone carrying payloads using multistatic radar[J]. The Journal of Engineering, 2019, 2019 (20): 7047.
22	Alvarez R, Borbor E, Grijalva F. Comparison of methods for signal analysis in the time-frequency domain[C]//IEEE 4th Ecuador Technical Chapters Meeting, 2019.
23	Xia J, Sun J, Yang H B, et al. Research on bark-frequency spectral coefficients heart sound classification algorithm based on multiple window time-frequency reassignment[J]. Journal of Biomedical Engineering, 2024, 41 (1): 51.
24	Balazs P, Dorfler M, Kowalski M, et al. Adapted and adaptive linear time-frequency representations: a synthesis point of view[J]. IEEE Signal Processing Magazine, 2013, 30 (6): 20. doi: 10.1109/msp.2013.2266075
25	Anastasiia K V, Ramo A, Tuomas H, et al. Controlled labelling of tracer antibodies for time-resolved fluorescence-based immunoassays[J]. Scientific Reports, 2024, 14 (1): 18113.
26	Elliot R W. Spontaneous subatomic mass-energy interconversion: implications for the heisenberg uncertainty principle and a theory of everything[J]. Applied Physics Research, 2025, 17 (1): 105. doi: 10.20944/preprints202501.1376.v1
27	Baird H W, Patton R K. The electromagnetic Doppler shift[J]. The Physics Teacher, 2025, 63 (6): 457. doi: 10.1119/5.0222222
28	Huang W, Guo H, LI X, et al. Researchon optimized YOLOv5s algorithm for detecting aircraft landing runway markings[J]. Processes, 2025, 13 (8): 2572.
29	Xu W F, Zhang H F, Zhang Y, et al. YOLO-DS: a detection model for desert shrub identification and coverage estimation in UAV remote sensing[J]. Journal of Forestry Research, 2025, 36 (1): 116. doi: 10.1007/s11676-025-01916-w
30	Guo S S, Wei J H, Chong J L, et al. Rockfall-YOLO deep learning model for rockfall motion object detection[J]. Journal of Computing in Civil Engineering, 2025, 39 (6): 04025104.
31	Zheng B S, Zhou J, Hong Z Q, et al. Vehicle recognition and driving information detection with uav video based on improved YOLOv5-Deep SORT algorithm[J]. Sensors, 2025, 25 (9): 2788. doi: 10.3390/s25092788
32	Huang S, Li X, Jiang Z, et al. Hyper feature fusion pyramid network for object detection [C]//IEEE International Conference on Multimedia & Expo Workshops, 2018.
33	Zhang N, Li J, Li Y R, et al. Global attention pyramid network for semantic segmentation[C]//Chinese Control Conference, 2019: 8728.

指标	时频分析算法
指标	FrFT	STFT	WVD	SPWVD	CWD	RSPWVD
运算/MFLOPs	2.3425	0.0047	6.5046	32.9844	54.6960	112.7376
执行时间/ms	1.1	8.7	16.9	70.3	95.8	262.5
内存使用/MB	1.05	1.13	2.97	5.98	5.23	29.23

参数	单旋翼双叶片	四旋翼双叶片	四旋翼三叶片
旋翼数	1	4	4
叶片数	2	2	3
叶片半径/m	0.13	0.12	0.067
转速/（r/s）	25	60	50

参数	参数值
射频频率/GHz	5.5
带宽/MHz	20
采样率/MHz	80
脉冲宽度/μs	20
脉冲重复间隔/μs	60
相干脉冲间隔/ms	192

指标	算法
指标	STFT	WVD	CWD	SPWVD	RSPWVD
$ {\sigma }_{{\mathrm{norm}}} $	0.084	0.084	0.115	0.056	0.041
$ {R}_{{\mathrm{tf}}} $	38.881	1.126	20.840	238.938	2.927

指标	算法
指标	STFT	WVD	CWD	SPWVD	RSPWVD
$ {\sigma }_{{\mathrm{norm}}} $	0.101	0.117	0.213	0.121	0.116
$ {R}_{{\mathrm{tf}}} $	206.700	8.611	122.759	382.570	17.270

基于RSPWVD重排谱图算法的无人机识别研究

UAV recognition research using RSPWVD spectrogram reassignment algorithm

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 17

参考文献 33

相关文章 6

编辑推荐

Metrics

本文评价

时频算法	指标
时频算法	P	R	mAP@0.5:0.95
STFT	0.995	0.978	0.725
SPWVD	0.958	0.978	0.839
RSPWVD	0.997	0.99	0.888

[1]	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.
[2]	杨磊, 孙卫天, 毛欣瑶, 夏亚波, 桑婧隺. 雷达鸟类目标微多普勒贝叶斯增强算法[J]. 系统工程与电子技术, 2024, 46(2): 505-516.
[3]	李海军, 孔繁程, 林云. 基于改进YOLOv5s的红外舰船检测算法[J]. 系统工程与电子技术, 2023, 45(8): 2415-2422.
[4]	成倩, 李佳, 杜娟. 基于YOLOv5的光学遥感图像舰船目标检测算法[J]. 系统工程与电子技术, 2023, 45(5): 1270-1276.
[5]	贺翥祯, 李敏, 苟瑶, 杨爱涛. 改进YOLOv5的合成孔径雷达图像舰船目标检测方法[J]. 系统工程与电子技术, 2023, 45(12): 3743-3753.
[6]	钱坤, 李晨瑄, 陈美杉, 王瑶. 基于YOLOv5的舰船目标及关键部位检测算法[J]. 系统工程与电子技术, 2022, 44(6): 1823-1832.

指标	算法
指标	STFT	WVD	CWD	SPWVD	RSPWVD
$ {\sigma }_{\text{norm}} $	0.084	0.099	0.189	0.110	0.104
$ {{R}}_{\text{tf}} $	263.964	365.131	189.721	115.143	2.098