基于代价敏感学习DBN-XGBoost的海面小目标检测方法研究

doi:10.12305/j.issn.1001-506X.2026.02.03

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (2): 402-409.doi: 10.12305/j.issn.1001-506X.2026.02.03

• 电子技术 • 上一篇

基于代价敏感学习DBN-XGBoost的海面小目标检测方法研究

缪燕子¹, 赵志非¹, 吴巍²^,*()

1. 中国矿业大学信息与控制工程学院，江苏徐州 221116
2. 海军工程大学兵器工程学院，湖北武汉 430033

收稿日期:2024-11-18 修回日期:2025-02-25 出版日期:2025-04-15 发布日期:2025-04-15
通讯作者: 吴巍 E-mail:wkw_wuwei@126.com
作者简介:缪燕子（1981—），女，教授，博士，主要研究方向为机器学习、智能机器人
赵志非（1998—），男，硕士研究生，主要研究方向为机器学习
基金资助:
国家自然科学基金（62473370）资助课题

Research on detection method of small sea surface targets based on cost-sensitive learning DBN-XGBoost

Yanzi MIAO¹, Zhifei ZHAO¹, Wei WU²^,*()

1. School of Information and Control Engineering，China University of Mining and Technology，Xuzhou 221116，China
2. School of Weapon Engineering，Naval University of Engineering，Wuhan 430033，China

Received:2024-11-18 Revised:2025-02-25 Online:2025-04-15 Published:2025-04-15
Contact: Wei WU E-mail:wkw_wuwei@126.com

摘要/Abstract

摘要：

针对海面小目标检测中特征提取能力弱、样本不平衡的难题，提出基于代价敏感学习的深度信念网络（deep belief network，DBN）与极端梯度提升（extreme gradient boosting，XGBoost）融合检测模型。通过DBN自动提取雷达回波深层特征，克服人工特征选择的主观性. 结合XGBoost分类器引入代价敏感矩阵，缓解海杂波与目标回波的数据不平衡。基于IPIX （Ice Multiparameter Imaging X-Band Radar）1993雷达数据集的测试表明，在0.512 s观测时间与0.001虚警率下，检测率较三特征检测器提升56%；在1.024 s观测时间条件下检测率达0.875，优于Hurst指数检测器、三特征检测器及快速自适应聚类梯度提升树方法。通过理论推导实现虚警率自适应控制机制，突破固定阈值限制。研究证明深度特征学习与代价敏感算法的协同优势，为海洋监测系统提供更优解决方案。

关键词: 海面小目标检测, 代价敏感学习, 不平衡数据集

Abstract:

To address the challenges of weak feature extraction and sample imbalance in maritime small sea surface targets detection, a cost-sensitive learning-based integrated detection model combining deep belief network （DBN） and extreme gradient boosting （XGBoost） is proposed. The DBN automatically extracts deep-level features from radar echo signals, overcoming the subjectivity of manual feature selection. The XGBoost classifier incorporates a cost-sensitive matrix to alleviate data imbalance between sea clutter dataset and target echoes. Tests on the Ice Multiparameter Imaging X-Band Radar 1993 radar dataset demonstrate that under 0.512 s observation time with 0.001 false alarm rate, the model improves detection rate by 56% compared to the three-feature detector; achieves 0.875 detection rate at observation time 1.024 s observation time, outperforming Hurst index detector, three-feature detector, and fast adaptive clustering gradient boosted decision trees method. Theoretical derivation enables adaptive false alarm rate control mechanism that overcomes the limitations of fixed thresholds. This study verifies the synergistic advantages of deep feature learning and cost-sensitive algorithms, providing an enhanced solution for marine monitoring systems.

Key words: sea surface small target detection, cost-sensitive learning, imbalanced datasets

中图分类号:

TP 391

缪燕子, 赵志非, 吴巍. 基于代价敏感学习DBN-XGBoost的海面小目标检测方法研究[J]. 系统工程与电子技术, 2026, 48(2): 402-409.

Yanzi MIAO, Zhifei ZHAO, Wei WU. Research on detection method of small sea surface targets based on cost-sensitive learning DBN-XGBoost[J]. Systems Engineering and Electronics, 2026, 48(2): 402-409.

图/表 8

图1

图2

图3

图4

图5

图6

图7

表1

参考文献 28

1	关键. 雷达海上目标特性综述[J]. 雷达学报, 2020, 9 (4): 674- 683.
	GUAN J. Summary of marine radar target characteristics[J]. Journal of Radars, 2020, 9 (4): 674- 683.
2	苏宁远, 陈小龙, 关键, 等. 基于深度学习的海上目标一维序列信号目标检测方法[J]. 信号处理, 2020, 36 (12): 1987- 1997.
	SU N Y, CHEN X L, GUAN J, et al. One dimensional sequence signal detection method for marine target based on deep learning[J]. Journal of Signal Processing, 2020, 36 (12): 1987- 1997.
3	张玉石, 许心瑜, 尹雅磊, 等. L 波段小擦地角海杂波幅度统计特性研究[J]. 电子与信息学报, 2014, 36 (5): 1044- 1048.
	ZHANG Y S, XU X Y, YIN Y L, et al. Research on amplitude statistical characteristics of L-band clutter corner sea clutter[J]. Journal of Radars, 2014, 36 (5): 1044- 1048.
4	SHUI P L, LI D C, XU S W. Tri-feature-based detection of floating small targets in sea clutter[J]. IEEE Trans. on Acrospace & Electronic Systems, 2014, 50(2): 1416−1430.
5	XU X K. Low observable targets detection by joint fractal properties of sea clutter: an experimental study of IPIX OHGR datasets[J]. IEEE Trans. on Antennas and Propagation, 2010, 58 (4): 1425- 1429. doi: 10.1109/TAP.2010.2041144
6	郭子薰, 水鹏朗, 白晓惠, 等. 海杂波中基于可控虚警K近邻的海面小目标检测[J]. 雷达学报, 2020, 9 (4): 654- 663. doi: 10.12000/JR20055
	GUO Z X, SHUI P L, BAI X H, et al. Sea-surface small target detection based on K-NN with controlled false alarm rate in sea clutter[J]. Journal of Radars, 2020, 9 (4): 654- 663. doi: 10.12000/JR20055
7	SHUI P L, LI D C. K-nearest neighbor classifier for sea-surface small target detection in sea clutter[J]. IEEE Trans. on Aerospace and Electronic Systems, 2011, 47 (4): 2416- 2426.
8	LI Y Z, XIE P C, TANG Z S, et al. SVM based sea-surface small target detection: a false-alarm-ratecontrollable approach[J]. IEEE Geoscience and Remote Sensing Letters, 2019, 16 (8): 1225- 1229. doi: 10.1109/LGRS.2019.2894385
9	陈世超, 高鹤婷, 罗丰. 基于极化联合特征的海面目标检测方法[J]. 雷达学报, 2020, 9 (4): 664- 673. doi: 10.12000/JR20072.
	CHEN S C, GAO H T, LUO F. Target detection in sea clutter based on combined characteristics of polarization[J]. Journal of Radars, 2020, 9 (4): 664- 673. doi: 10.12000/JR20072.
10	李骁, 施赛楠, 董泽远, 等. 基于时频域深度网络的海面小目标特征检测[J]. 雷达科学与技术, 2022, 20 (2): 209- 216，230. doi: 10.3969/j.issn.1672-2337.2022.02.013
	LI X, SHI S N, DONG Z Y, et al. Feature detection of small sea-surface target via deep network in time-frequency domain[J]. Radar Science and Technology, 2022, 20 (2): 209- 216，230. doi: 10.3969/j.issn.1672-2337.2022.02.013
11	何友, 关键, 孟祥伟, 等. 雷达目标检测与恒虚警处理[M]. 第2版. 北京: 清华大学出版社, 2011.
	HE Y, GUAN J, MEND X W, et al. Radar target detection and CFAR proccssing[M]. 2nd ed. Beijing: Tsingliua University Press, 2011.
12	施赛楠, 杨静, 王杰. 基于多域多维特征融合的海面小目标检测[J]. 信号处理, 2020, 36 (12): 2099- 106.
	SHI S N, YANG J, WANG J. Detection of sea surface small target based on multi-domain and multi-dimensional feature fusion[J]. Journal of Signal Processing, 2020, 36 (12): 2099- 106.
13	薛春岭, 曹菲, 孙庆, 等. 基于多特征信息融合的海面微弱目标检测[J]. 系统工程与电子技术, 2022, 44 (11): 3338- 3345. doi: 10.12305/j.issn.1001-506X.2022.11.07
	XUE C L, CAO F, SUN Q, et al. Sear surface weak target detection based on multi-feature information fusion[J]. Systems Engineering and Electronics, 2022, 44 (11): 3338- 3345. doi: 10.12305/j.issn.1001-506X.2022.11.07
14	GUO Z X, BAI X H, LI J Y, et al. Fast detection of small targets in high-resolution maritime radars by feature normalization and fusion[J]. IEEE Journal of Oceanic Engineering, 2022, 47 (3): 736- 750. doi: 10.1109/JOE.2021.3133553
15	施赛楠, 杨静, 董泽远. 基于高维特征域随机森林的海面小目标检测[J]. 现代雷达, 2022, 44 (3): 63- 69.
	SHI S N, YANG J, DONG Z Y. Detection of small sea-surface target based on random forest in high-dimensional feature domain[J]. Modern Radar, 2022, 44 (3): 63- 69.
16	薛安克, 毛克成, 张乐. 多分类器联合虚警可控的海上小目标检测方法[J]. 电子与信息学报, 2023, 45 (7): 2528- 2536. doi: 10.11999/JEIT220710
	XUE A K, MAO K C, ZHANG L. Multi-feature marine small target detection based on multi-class classifier[J]. Journal of Electronics & Information Technology, 2023, 45 (7): 2528- 2536. doi: 10.11999/JEIT220710
17	时艳玲, 刘子鹏, 贾邦玲. 样本不平衡下的海杂波弱目标分类研究[J]. 信号处理, 2021, 37 (9): 1781- 1789.
	SHI Y L, LIU Z P, JIA B L. The unbalanced classification of weak target in sea clutter[J]. Journal of Signal Processing, 2021, 37 (9): 1781- 1789.
18	尹翔, 李婉华. 基于GAN目标数据增强的海面小目标检测[J]. 现代雷达, 2024, 47 (4): 20- 28.
	YIN X, LI W H. Sea-surface small target detection based on target data augmentation using GAN[J]. Modern Radar, 2024, 47 (4): 20- 28.
19	李京泰, 王晓丹. 基于代价敏感激活函数XGBoost的不平衡数据分类方法[J]. 计算机科学, 2022, 49 (5): 135- 143.
	LI J T, WANG X D. XGBoost for imbalanced data based on cost-sensitive activation function[J]. Computer Science, 2022, 49 (5): 135- 143.
20	阎妍. 海杂波中的噪声抑制及其微弱信号检测[D]. 南京: 南京信息工程大学, 2022.
	YAN Y. Noise suppression and weak signal detection in sea clutter[D]. Nanjing: Nanjing University of Information Science and Technology, 2022.
21	HU J, TUNG W W, GAO J B. Detection of low observable targets within sea clutter by structure function based multifractal analysis[J]. IEEE Trans. on Antennas and Propagation, 2006, 54(1): 136−143.
22	SHI S, SHUI P. Sea-surface floating small target detection by one class classifier in time-frequency feature space[J]. IEEE Trans. on Geoscience and Remote Sensing, 2018, 56, 6395- 6411. doi: 10.1109/TGRS.2018.2838260
23	杨静. 高维特征域海面小目标检测方法研究[D]. 南京: 南京信息工程大学, 2022.
	YANG J. Research on detection methods of small sea surface targets in high dimensional feature domain[D]. Nanjing: Nanjing University of Information Science and Technology, 2022.
24	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 & Electronic Systems, 2014, 50(2): 1416−1430.
25	GUO Z X, SHUI P L. Anomaly based sea-surface small target detection using K-nearest neighbor classification[J]. IEEE Trans. on Aerospace and Electronic Systems, 2020, 56 (6): 4947- 4964.
26	HINTON G E. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7): 1527−1554.
27	HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5): 504−507.
28	HOCHBERG E. A new generation of radar target signature data[J]. IEEE Trans. on Aerospace and Electronic Systems, 2006, 29 (1): 178- 184.

[1]	唐玺博, 张立民, 钟兆根. 基于ADASYN与改进残差网络的入侵流量检测识别[J]. 系统工程与电子技术, 2022, 44(12): 3850-3862.
[2]	田江，顾宏. 一种基于混合策略的孤立点检测方法[J]. Journal of Systems Engineering and Electronics, 2010, 32(8): 1775-1779.

基于代价敏感学习DBN-XGBoost的海面小目标检测方法研究

Research on detection method of small sea surface targets based on cost-sensitive learning DBN-XGBoost

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 28

相关文章 2

编辑推荐

Metrics

本文评价

检测器	观测时间/s	虚警率	检测率
三特征	1.024	0.001	0.618
时频三特征	1.024	0.001	0.732
K-NN	1.024	0.001	0.812
所提检测器	1.024	0.001	0.875