基于目标时频起伏特性的信号重构与检测方法

doi:10.12305/j.issn.1001-506X.2026.09.12

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (9): 2987-2998.doi: 10.12305/j.issn.1001-506X.2026.09.12

• 传感器与信号处理 • 上一篇

基于目标时频起伏特性的信号重构与检测方法

杨翰琨¹(), 陈宝欣², 刘志坤³(), 刘宁波², 薛伟¹

1. 哈尔滨工程大学烟台研究院，山东烟台 265500
2. 海军航空大学信息融合研究所，山东烟台 264001
3. 海军工程大学，湖北武汉 430033

收稿日期:2025-09-18 修回日期:2025-12-17 接受日期:2026-02-04 出版日期:2026-03-26 发布日期:2026-03-26
通讯作者: 刘志坤 E-mail:yanghankun@hrbeu.edu.cn;2034006859@qq.com
作者简介:杨翰琨（2001—），男，硕士研究生，主要研究方向为雷达目标检测与杂波抑制
陈宝欣（1990—），男，工程师，博士，主要研究方向为阵列信号处理、雷达多维信号处理
刘宁波（1983—），男，教授，博士，主要研究方向为雷达信号智能处理、海上目标探测
薛　伟（1970—），男，教授，博士，主要研究方向为水下及地下无线通信技术、通信信号检测与识别技术
基金资助:
国家自然科学基金（62388102，62101583）；泰山学者工程（tsqn202211246）资助课题

Signal reconstruction and detection method based on target time-frequency fluctuation characteristic

Hankun Yang¹(), Baoxin Chen², Zhikun Liu³(), Ningbo Liu², Wei Xue¹

1. Yantai Research Institute，Harbin Engineering University，Yantai 265500，China
2. Information Fusion Institute，Navy Aviation University，Yantai 264001，China
3. Naval University of Engineering，Wuhan 430033，China

Received:2025-09-18 Revised:2025-12-17 Accepted:2026-02-04 Online:2026-03-26 Published:2026-03-26
Contact: Zhikun Liu E-mail:yanghankun@hrbeu.edu.cn;2034006859@qq.com

摘要/Abstract

摘要：

针对强杂波背景下海面运动目标难以检测的问题，提出一种基于相关模态重构的特征检测方法。首先，通过深度模型对目标时频谱进行智能分割，构建理想目标参考模板信号。进而，在变分模态分解框架下，设计多维相关性指标评分策略，评估各模态信号与模板信号之间的匹配程度，从而精确筛选目标模态。在完成模态重构之后，进一步对重构信号进行多域特征提取与筛选，并进行非线性特征融合。实验结果表明，在每种特征检测方法中，重构信号的检测概率均显著高于原始信号。融合特征检测方法在单特征检测基础上进一步提升了检测概率。

Abstract:

To address the challenge of detecting sea-surface moving targets in strong clutter environments, a feature detection method based on correlational mode reconstruction is proposed. The method first uses a deep learning model to intelligently segment the time-frequency spectrogram of the target, creating an ideal target reference model signal. Subsequently, within the framework of variational mode decomposition, a multi-dimensional correlation metric scoring strategy is designed to evaluate the matching degree between each mode signal and the model signal, which enables the precise selection of target modes. After mode reconstruction is completed, multi-domain features are further extracted and selected, followed by non-linear feature fusion. Experimental results show that the detection probability of the reconstructed signal is significantly higher than that of the original signal in every feature detection method. Furthermore, the fused feature detection method provides an additional performance boost over single-feature detection.

Key words: correlational mode reconstruction, sea clutter suppression, radar target detection, time-frequency spectrum segmentation, kernel discriminant analysis

中图分类号:

TN 957.51

杨翰琨, 陈宝欣, 刘志坤, 刘宁波, 薛伟. 基于目标时频起伏特性的信号重构与检测方法[J]. 系统工程与电子技术, 2026, 48(9): 2987-2998.

Hankun Yang, Baoxin Chen, Zhikun Liu, Ningbo Liu, Wei Xue. Signal reconstruction and detection method based on target time-frequency fluctuation characteristic[J]. Systems Engineering and Electronics, 2026, 48(9): 2987-2998.

图/表 14

图1

图2

图3

图4

表1

表2

表3

表4

表5

表6

图5

图6

表7

表8

参考文献 30

1	王威, 杨勇, 韩静雯. 基于相关性的雷达海面弱目标检测技术[J/OL]. 系统工程与电子技术, [2025-09-11]. https: //link.cnki.net/urlid/11.2422.TN.20250320.1357.002.
2	Wen B T, Wei Y B, Lu Z Z. Sea clutter suppression and target detection algorithm of marine radar image sequence based on spatio-temporal domain joint filtering[J]. Entropy, 2022, 24 (2): 250. doi: 10.3390/e24020250
3	Huang P H, Yang H, Xia X G, et al. A novel sea clutter rejection algorithm for spaceborne multichannel radar systems[J]. IEEE Trans. on Geoscience and Remote Sensing, 2022, 60, 5117422. doi: 10.1109/tgrs.2022.3204324
4	任晓昱, 林瑞奇, 邓云开, 等. 基于模态分解的地基雷达慢速弱目标检测方法[J]. 系统工程与电子技术, 2025, 47 (1): 70. doi: 10.12305/j.issn.1001-506X.2025.01.08
5	Zhang K Y, He W, Liu L N, et al. Sea clutter denoising based on complete ensemble empirical mode decomoposition[C]//4th International Conference on Natural Language Processing, 2022: 282.
6	李波, 胡哿郗, 石剑钧, 等. 基于多惩罚因子优化VMD的滚动轴承故障特征提取方法[J]. 系统工程与电子技术, 2023, 45 (11): 3690. doi: 10.12305/j.issn.1001-506X.2023.11.38
7	Sun J, Xing H Y, Wu J J. Distributed sea clutter denoising algorithm based on variational mode decomposition[J]. Instrumentation, 2020, 7 (3): 23.
8	Liu N B, Yang H K, Wang G Q, et al. A spectral mode reconstruction method for floating target detection under strong sea clutter conditions[J]. Remote Sensing, 2025, 17 (18): 3155. doi: 10.3390/rs17183155
9	Li Z, Zhao Y, Hu X Y, et al. ECOD: unsupervised outlier detection using empirical cumulative distribution functions[J]. IEEE Trans. on Knowledge and Data Engineering, 2022, 35 (12): 12181. doi: 10.2139/ssrn.4313179
10	Laasch N, Braun W, Knoff L, et al. Comparison of derivative-based and correlation-based methods to estimate effective connectivity in neural networks[J]. Scientific Reports, 2025, 15 (1): 5357. doi: 10.1038/s41598-025-88596-y
11	Miao Y H, Zhang B Y, Li C H, et al. Feature mode decomposition: new decomposition theory for rotating machinery fault diagnosis[J]. IEEE Trans. on Industrial Electronics, 2022, 70 (2): 1949.
12	Lu Y, Wu Y, Liu B, et al. Cross-modality person re-identification with shared-specific feature transfer[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 13379.
13	Cong H S, Sun J F, Lu Z Y, et al. Intrapulse coherence degradation suppressing method of echo signal in coherent lidar[J]. Chinese Optics Letters, 2023, 21 (7): 071201. doi: 10.3788/COL202321.071201
14	Wang Y Q, Zhao Y, Petzold L. An empirical study on the robustness of the segment anything model （SAM）[J]. Pattern Recognition, 2024, 155, 110685. doi: 10.1016/j.patcog.2024.110685
15	Xie J D, Xu X J. Phase-feature-based detection of small targets in sea clutter[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 19, 3507405. doi: 10.1109/lgrs.2021.3093620
16	Guo Z X, Shui P L, Bai X H. Small target detection in sea clutter using all-dimensional Hurst exponents of complex time sequence[J]. Digital Signal Processing, 2020, 101, 102707. doi: 10.1016/j.dsp.2020.102707
17	Adebiyi M O, Arowolo M O, Mshelia M D, et al. A linear discriminant analysis and classification model for breast cancer diagnosis[J]. Applied Sciences, 2022, 12 (22): 11455. doi: 10.3390/app122211455
18	Bai X, Zhang M X, Jin Z H, et al. Fault detection and diagnosis for chiller based on feature-recognition model and kernel discriminant analysis[J]. Sustainable Cities and Society, 2022, 79, 103708. doi: 10.1016/j.scs.2022.103708
19	Shui P L, Li D C, Xu S W. Tri-feature-based detection of floating small targets in sea clutter[J]. IEEE Trans. on Aerospace and Electronic Systems, 2014, 50 (2): 1416. doi: 10.1109/TAES.2014.120657
20	Wu X J, Ding H, Liu N B, et al. A method for detecting small targets in sea surface based on ridges-radon transform[J]. Journal of Signal Processing, 2021, 37 (9): 1599.
21	Yu W Y, Yang G Y, Zhang H L. A new method for ship length estimation based on Doppler spectrum analysis[J]. Radar Science and Technology, 2015, 13 (5): 522.
22	Guan J, Jiang X Y, Liu N B, et al. Target detection in sea clutter based on feature re-expression using Spearman correlation[J]. IEEE Sensors Journal, 2024, 24 (19): 30435. doi: 10.1109/JSEN.2024.3440065
23	华宇辉, 胡文军, 崔胜男, 等. 基于特征重要性的质量相关故障诊断与量化评估[J]. 测控技术, 2025, 44 (4): 25. doi: 10.19708/j.ckjs.2025.02.211
24	Cai W, Huang Y, Davidson T N. Efficient transceiver design for MIMO dual-function radar-communication systems[J]. IEEE Trans. on Signal Processing, 2023, 71, 1786. doi: 10.1109/TSP.2023.3275274
25	Cai W, Huang Y, Zheng L, et al. Transmit waveform design for dual-function radar-communication systems via hybrid linear-nonlinear precoding[J]. IEEE Trans. on Signal Processing, 2023, 71, 2130. doi: 10.1109/TSP.2023.3278858
26	Cai W, Huang Y, Peng J Y, et al. Slow-time FDA-MIMO technique with application to STAP radar[J]. IEEE Trans. on Aerospace and Electronic Systems, 2022, 58 (1): 74. doi: 10.1109/TAES.2021.3098100
27	刘宁波, 李佳, 王国庆, 等. 雷达对海探测试验与目标特性数据获取——海上目标多源观测数据集[J]. 雷达学报（中英文）, 2025, 14 (3): 754. doi: 10.12000/JR25001.
28	关键, 刘宁波, 王国庆, 等. 雷达对海探测试验与目标特性数据获取——海上目标双极化多海况散射特性数据集[J]. 雷达学报, 2023, 12 (2): 456. doi: 10.12000/JR23029
29	刘宁波, 丁昊, 黄勇, 等. X波段雷达对海探测试验与数据获取年度进展[J]. 雷达学报, 2021, 10 (1): 173.
30	刘宁波, 董云龙, 王国庆, 等. X波段雷达对海探测试验与数据获取[J]. 雷达学报, 2019, 8 (5): 656.

相关性指标	取值范围	判别准则	强目标	弱目标	杂波
互相关	[−1,1]	越接近1越相关	0.792 7	−0.247 7	−0.010 3
PCC	[−1,1]	越接近1越相关	0.667 2	0.088 7	−0.043 0
MI	[0,+∞]	数值越大越相关	0.396 8	0.040 0	0.024 0
余弦相似度	[−1,1]	越接近1越相关	0.792 7	−0.247 7	−0.005 4
频谱角度相似度	[0,π/2]	越接近0越相关	0.333 1	0.523 5	0.517 1
CPS	[0,+∞]	数值越大越相关	0.024 8	0.008 3	0.003 2

参数/信息	具体值/说明
工作模式	凝视
工作频率/GHz	9.3 ~ 9.5
发射时间	40 ns ~ 100 µs
发射功率/W	100
距离分辨率/m	6
采集持续时间/s	65.5
脉冲重复频率/Hz	2000

数据组别	海况等级	有效浪高	浪向	距离维采样总点数	脉冲总数	目标距离单元范围	强目标距离单元	弱目标距离单元	对照杂波距离单元
1	5级	2.7 m	北西北	1 000	67 584	663~674	669	674	689
2	4级	2.3 m	北东北	1 000	131 072	661~672	667	672	687

IMF	中心频率/Hz	CPS	PCC	MI	综合得分	目标模态
1	0.00	0.029 59	−0.154 10	1.993 12	0.751 28	是
2	27.36	0.018 88	0.369 95	1.715 88	0.643 94	是
3	74.21	0.000 43	0.080 34	1.741 77	0.367 28	否
4	82.03	0.000 52	−0.122 26	1.700 56	0.291 61	否
5	109.38	0.000 18	0.058 14	1.483 09	0.233 82	否
6	128.92	0.000 45	−0.456 08	1.686 44	0.190 55	否
7	171.84	0.000 13	0.076 44	1.851 30	0.415 43	否
8	195.35	0.000 15	−0.224 25	1.676 57	0.247 37	否
9	234.36	0.000 08	−0.183 66	1.749 31	0.292 95	否
10	273.44	0.000 06	0.203 06	1.697 17	0.376 04	否
11	347.65	0.000 08	0.068 85	1.611 78	0.297 58	否
12	382.81	0.000 13	−0.083 64	1.804 87	0.348 22	否
13	445.32	0.000 06	−0.170 38	1.742 26	0.292 99	否
14	511.72	0.000 06	0.321 74	1.728 93	0.424 51	否
15	593.75	0.000 07	−0.411 85	1.459 56	0.089 52	否
16	726.63	0.000 09	−0.197 35	1.533 76	0.185 49	否
17	828.26	0.000 06	0.144 48	1.490 49	0.260 22	否
18	819.98	0.000 05	0.136 91	1.495 18	0.260 16	否
19	−886.70	0.000 05	−0.428 13	1.492 02	0.100 34	否
20	−757.82	0.000 06	0.104 18	1.496 33	0.251 65	否
21	−628.96	0.000 07	−0.097 79	1.584 13	0.237 47	否
22	−488.25	0.000 03	0.325 38	1.300 08	0.218 97	否
23	−296.89	0.000 15	−0.050 97	1.507 73	0.214 76	否
24	−117.18	0.000 38	0.565 22	1.331 50	0.305 25	否
25	−15.64	0.023 99	0.733 52	1.654 19	0.773 81	是

特征	重构信号特征重要性	Fisher J值
特征	重构信号特征重要性	原始信号	重构信号
DPH	0.488 9	1.383 0	1.458 3
AA	0.225 2	1.466 3	1.492 2
SC	0.086 1	0.330 3	1.043 9
RRT-MV	0.028 5	1.061 8	1.319 3
PH	0.023 6	1.708 1	1.644 4
RI	0.022 3	0.789 0	0.531 5
PE	0.018 1	0.438 4	0.709 4
MS	0.013 2	0.978 3	0.329 5
FEM	0.010 1	0.463 4	0.009 4
NR	0.009 5	0.898 2	0.353 5
AE	0.009 2	0.878 2	0.001 6
SD	0.008 9	0.760 1	0.624 4
SK	0.008 7	0.003 7	0.031 2
SOFE	0.008 6	0.001 2	0.021 7
CCC	0.008 2	0.652 9	0.447 8
TEM	0.007 5	0.649 4	0.655 5
RRT-BW	0.007 1	0.077 6	0.377 4
SOTE	0.006 8	0.015 8	0.048 1
SBW	0.006 6	1.514 7	0.000 7
VF	0.003 1	0.149 3	0.181 4
MDP	0.000 0	0.000 0	0.000 0

基于目标时频起伏特性的信号重构与检测方法

Signal reconstruction and detection method based on target time-frequency fluctuation characteristic

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 30

相关文章 12

编辑推荐

Metrics

本文评价

特征	目标均值	杂波均值	目标方差	杂波方差	类内散度	类间散度	Fisher J
AA	8.91×10⁴	1.52×10⁴	3.45×10⁹	2.07×10⁸	5.46×10⁹	3.66×10⁹	1.492 2
PH	1.16×10⁵	2.49×10⁴	4.56×10⁹	5.02×10⁸	8.32×10⁹	6.06×10⁹	1.644 4
DPH	2.43×10⁶	3.91×10⁵	2.69×10¹²	1.60×10¹¹	4.15×10¹²	2.85×10¹²	1.458 3
RRT-MV	1.82×10¹	1.05×10¹	2.40×10¹	2.10×10¹	5.94×10¹	4.50×10¹	1.319 3
SC	−3.73	7.53×10¹	2.15×10³	3.84×10³	6.25×10³	5.99×10³	1.043 9
LDA融合特征	8.30×10⁻⁴	−8.30×10⁻⁴	4.67×10⁻⁷	3.19×10⁻⁷	2.76×10⁻⁶	7.86×10⁻⁷	3.505 7
KDA融合特征	2.75×10⁻³	−2.75×10⁻³	1.57×10⁻⁶	1.00×10⁻⁶	3.04×10⁻⁵	2.58×10⁻⁶	11.778 4

[1]	王威, 杨勇, 韩静雯. 基于相关性的雷达海面弱目标检测技术[J]. 系统工程与电子技术, 2026, 48(3): 787-794.
[2]	吴巍, 薛冰, 刘丹丹. 基于Tri-feature训练的目标与海杂波鉴别算法[J]. 系统工程与电子技术, 2024, 46(9): 2935-2940.
[3]	张俊玲, 董玫, 陈伯孝. 基于可调Q因子小波变换的海杂波抑制算法[J]. 系统工程与电子技术, 2023, 45(2): 343-351.
[4]	杨威, 刘永祥, 付耀文, 黎湘. 基于连续信任函数理论的鲁棒雷达目标检测方法[J]. 系统工程与电子技术, 2021, 43(9): 2463-2469.
[5]	董自巍, 孙俊, 孙晶明, 潘美艳. 稀疏字典学习海面微弱动目标检测[J]. 系统工程与电子技术, 2020, 42(1): 30-36.
[6]	时艳玲, 林毓峰, 梁丹丹. 非平稳海杂波背景下子带分段ANMF检测器[J]. 系统工程与电子技术, 2018, 40(4): 782-789.
[7]	曹健, 王兆祎, 胡进峰, 何子述. 基于知识辅助的天波雷达海杂波抑制方法[J]. 系统工程与电子技术, 2018, 40(3): 533-537.
[8]	李会勇，许丁文，胡进峰，戴文娟，何子述. 基于FRFT的天波雷达多机动目标检测[J]. 系统工程与电子技术, 2014, 36(9): 1725-1730.
[9]	时艳玲. 空间非均匀海杂波的协方差矩阵估计新算法[J]. 系统工程与电子技术, 2013, 35(11): 2239-2243.
[10]	吴兆平，吴惠阳，苏涛. 距离-多普勒-时间三维雷达数据中的目标能量积累[J]. Journal of Systems Engineering and Electronics, 2013, 35(1): 68-72.
[11]	赵志国，陈建文，鲍拯. 一种改进的OTHR自适应海杂波抑制方法[J]. Journal of Systems Engineering and Electronics, 2012, 34(5): 909-914.
[12]	冯讯，王首勇，朱晓波. 基于SG-Alpha稳定分布模型的雷达目标Rao检测方法[J]. Journal of Systems Engineering and Electronics, 2012, 34(2): 287-291.

特征	原始信号	重构信号
AA	0.295 0	0.413 1
PH	0.261 5	0.359 3
DPH	0.293 2	0.416 0
RRT-MV	0.242 0	0.302 7
SC	0.113 5	0.265 3
LDA融合特征	0.431 9	0.553 3
KDA融合特征	0.529 4	0.635 9

特征	原始信号	重构信号
AA	0.4521	0.5436
PH	0.4412	0.5241
DPH	0.4365	0.5403
RRT-MV	0.2773	0.3315
SC	0.0445	0.2117
LDA融合特征	0.5093	0.5995
KDA融合特征	0.5992	0.7112