Improved YOLOv5-based radar object detection

doi:10.23919/JSEE.2025.000004

Journal of Systems Engineering and Electronics ›› 2025, Vol. 36 ›› Issue (4): 932-939.doi: 10.23919/JSEE.2025.000004

• • 上一篇

收稿日期:2023-09-28 出版日期:2025-08-18 发布日期:2025-09-04

Improved YOLOv5-based radar object detection

Zhicheng WANG¹^,²(), Weilin LI³(), Xiaoyi SUN²(), Hanxi ZHAO²(), Wentong CHEN²(), Jing WU³^,*()

¹ Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
² Research & Development Department, Shanghai Radio Equipment Research Institute, Shanghai 201109, China
³ Department of Automation and Key Laboratory of System Control and Information Processing, Shanghai Jiao Tong University, Shanghai 200240, China

Received:2023-09-28 Online:2025-08-18 Published:2025-09-04
Contact: Jing WU E-mail:wangzhicheng@sjtu.edu.cn;dan-stevens@sjtu.edu.cn;sunxiaoyi0711@163.com;1229748798@qq.com;706397074@139.com;jingwu@sjtu.edu.cn
About author:
WANG Zhicheng was born in 1982. He received his Ph.D. degree at the School of Electronic Information and Electrical Engineering, from Shanghai Jiao Tong University in 2024. He currently works as a researcher at Shanghai Radio Equipment Research Institute. His research interests are radar signal processing, target detection and recognition, and artificial intelligence. E-mail: wangzhicheng@sjtu.edu.cn

LI Weilin was born in 2000. He received his B.S. degree from Shanghai Jiao Tong University in 2022. He is pursuing his M.S. degree at the School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University. His research interests include artificial intelligence and image processing. E-mail: dan-stevens@sjtu.edu.cn

SUN Xiaoyi was born in 1998. She received her M.S. degree in the College of Electronic Science and Technology from National University of Defense Technology. She currently works as an assistant engineer at Shanghai Radio Equipment Research Institute. Her research interests are radar signal processing and image processing. E-mail: sunxiaoyi0711@163.com

ZHAO Hanxi was born in 1992. She received her M.S. degree in the School of Electrical and Electronic Engineering from Xidian University in 2017. Since 2017, she has been with Shanghai Radio Equipment Research Institute. Her research interests are radar signal processing and image processing. E-mail: 1229748798@qq.com

CHEN Wentong was born in 1979. He received his Ph.D. degree in the College of Electronic Science and Technology from National University of Defense Technology. His research interests are spatial situation awareness and early warning. Email: 706397074@139.com

WU Jing was born in 1979. She received her Ph.D. degree in electrical engineering from the University of Alberta, Edmonton, AB, Canada, in 2008. Since 2011, she has been with Shanghai Jiao Tong University, Shanghai, China, and is currently a professor. She is a registered Professional Engineer in Alberta, Canada. Her research interests include robust control, model predictive control and state estimation for cyber-physical systems. E-mail: jingwu@sjtu.edu.cn
Supported by:
This work was supported by the National Natural Science Foundation of China (62273236; 62136006; 62073215), and Key R&D Projects in Hainan Province (ZDYF2024GXJS009).

摘要/Abstract

Abstract:

In this paper, we propose an improved YOLOv5-based object detection method for radar images, which have the characteristics of diffuse weak noise and imaging distortion. To mitigate the effects of noise without losing spatial information, an coordinate attention (CA) has been added to pre-extract the feature of the images, which can guarantee a better feature extraction ability. A new stochastic weighted average (SWA) method is designed to refine generalization ability of the algorithm, where the medium mean is used instead of their average value. By introducing an deformable convolution, both regular and irregular images can be proceeded. The experimental results show that the improved algorithm performs better in object detection of radar images compared with the YOLOv5 model, which confirms the effectiveness and feasibility of our model.

Key words: YOLOv5, coordinate attention (CA), deformable convolution, radar image

. [J]. Journal of Systems Engineering and Electronics, 2025, 36(4): 932-939.

Zhicheng WANG, Weilin LI, Xiaoyi SUN, Hanxi ZHAO, Wentong CHEN, Jing WU. Improved YOLOv5-based radar object detection[J]. Journal of Systems Engineering and Electronics, 2025, 36(4): 932-939.

图/表 10

参考文献 31

1	LIU A F, ZHANG S H, ZHANG C, et al RaNeRF: neural 3D reconstruction of space targets from ISAR image sequences. IEEE Trans. on Geoscience and Remote Sensing, 2023, 61, 5107215.
2	ZHOU X Y, WANG Y, LU X F Approach for ISAR imaging of near-field targets based on coordinate conversion and image interpolation. Journal of Systems Engineering and Electronics, 2021, 32 (2): 425- 436. doi: 10.23919/JSEE.2021.000036
3	JIN X F, SU F L, LI H X, et al Automatic ISAR ship detection using triangle-points affine transform reconstruction algorithm. Remote Sensing, 2023, 15 (10): 2507. doi: 10.3390/rs15102507
4	YANG C L, ZHANG C, FAN Z L, et al. A multi-channel aggregation framework for object detection in large-scale sar image. Proc. of the IEEE International Conference on Acoustics, Speech and Signal Processing, 2023. DOI: 10.1109/ICASSP49357.2023.10096182.
5	ZHANG C Q, ZHANG Z W, CHONG M Z, et al. A joint unsupervised super-resolution and sar object detection network. Proc. of the 8th International Conference on Computer and Communication System, 2023: 1027–1030.
6	GUI S L, YANG Y, HU R Z, et al Dynamic ISAR imaging method for multiple moving vehicles based on OMP-CADMM. IEEE Trans. on Vehicular Technology, 2022, 71 (10): 10948- 10959. doi: 10.1109/TVT.2022.3183796
7	BARBARY M, HAFEZ A S, ABD El-AZEEM M H, et al Extended drones tracking from ISAR images with Doppler effect and orientation based robust sub-random matrices algorithm. IEEE Trans. on Vehicular Technology, 2022, 71 (12): 12648- 12666. doi: 10.1109/TVT.2022.3195810
8	YU C P, XIONG W, LI X Q, et al Deep convolutional neural network for meteorology target detection in airborne weather radar images. Journal of Systems Engineering and Electronics, 2023, 34 (5): 1147- 1157.
9	GONG Y C, ZHANG Z, WEN J B, et al Small ship detection of SAR images based on optimized feature pyramid and sample augmentation. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023, 16, 7385- 7392.
10	NOVAK L M, OWIRKA G J, BROWER W S, et al The automatic target-recognition system in SAIP. Lincoln Laboratory Journal, 1997, 10 (2): 187- 202.
11	SEUNG H S, LEE D D The manifold ways of perception. Science, 2000, 290 (5500): 2268- 2269. doi: 10.1126/science.290.5500.2268
12	ANAGNOSTOPOULOS G C. SVM-based target recognition from synthetic aperture radar images using target region outline descriptors. Nonlinear Analysis: Theory, Methods & Applications, 2009, 71(12): e2934-e2939.
13	AMOON M, REZAIRAD G A Automatic target recognition of synthetic aperture radar (SAR) images based on optimal selection of Zernike moments features. IET Computer Vision, 2013, 8 (2): 77- 85.
14	KEYDEL E, LEE S. Signature prediction for model-based automatic target recognition. Proc. of the SPIE-Algorithms for Synthetic Aperture Radar Imagery III, 1996: 306–317.
15	OLIVA A, TORRALBA A The role of context in object recognition. Trends in Cognitive Sciences, 2007, 20 (11): 520- 527.
16	NING W, CHEN W G, ZHANG X G. Automatic target recognition of ISAR object images based on neural network. Proc. of the IEEE International Conference on Neural Networks & Signal Processing, 2004: 373–376.
17	KARINE A, TOUMI A, KHENCHAF A, et al Target recognition in radar images using weighted statistical dictionary-based sparse representation. IEEE Geoscience and Remote Sensing Letters, 2017, 14 (12): 2403- 2407. doi: 10.1109/LGRS.2017.2766225
18	LEE S, LEE M, KIM K, et al Classification of ISAR images using variable cross-range resolutions. IEEE Trans. on Aerospace and Electronic Systems, 2018, 54 (5): 2291- 2303. doi: 10.1109/TAES.2018.2814211
19	ZOU Q, NI L H, ZHANG T, et al Deep learning based feature selection for remote sensing scene classification. IEEE Geoscience and Remote Sensing Letters, 2015, 11 (12): 2321- 2325.
20	LUO Y S, XU J H, FENG C X, et al An accurate detection algorithm for time backtracked projectile-induced water columns based on the improved YOLO network. Journal of Systems Engineering and Electronics, 2023, 34 (4): 981- 991.
21	WANG Q, YUAN Z, DU Q, et al GETNET: a general end-to-end 2-D CNN framework for hyperspectral image change detection. IEEE Trans. on Geoscience and Remote Sensing, 2019, 57 (1): 3- 13. doi: 10.1109/TGRS.2018.2849692
22	ZAIED S, TOUMI A, KHENCHAF A. Target classification using convolutional deep learning and auto-encoder models. Proc. of the International Conference on Advanced Technologies for Signal and Image Processing, 2018. DOI: 10.1109/ATSIP.2018.8364502.
23	DENG S, DU L, LI C, et al SAR automatic target recognition based on Euclidean distance restricted autoencoder. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 7 (10): 3323- 3333.
24	HUANG Z L, PAN Z X, LEI B Transfer learning with deep convolutional neural network for SAR target classification with limited labeled data. Remote Sensing, 2017, 9 (9): 907.
25	PEI J, HUANG Y L, HUO W B, et al SAR automatic target recognition based on multiview deep learning framework. IEEE Trans. on Geoscience and Remote Sensing, 2018, 4 (56): 2196- 2210.
26	HOU Q B, ZHOU D Q, FENG J S. Coordinate. attention for efficient mobile network design. Proc. of the Conference on Computer Vision and Pattern Recognition, 2021: 13708−13717.
27	CALDAROLA D, CAPUTO B, CICCONE M. Improving generalization in federated learning by seeking flat minima. Proc. of the European Conference on Computer Vision, 2022: 654−672.
28	REN S, HE K, GIRSHICK R, et al Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2017, 39 (6): 1137- 1149. doi: 10.1109/TPAMI.2016.2577031
29	CALDAROLA D, CAPUTO B, CICCONE M. Improving generalization in federated learning by seeking flat minima. Proc. of the European Conference on Computer Vision, 2022: 654–672.
30	HU J, SHEN L, SUN G. Squeeze and excitation networks. Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7132–7141.
31	WOO S, PARK J, LEE J Y, et al. Cbam: convolutional block attention module. Proc. of the European Conference on Computer Vision, 2018: 3–19.

Environment	Configuration
System	Windows10
CPU	AMD Ryzen 7 4800H with Radeon Graphics
GPU	RTX 2060
Pytorch	1.10.0

Model	mAP_0.5
SSD	0.7215
Faster R-CNN	0.8066
YOLOv5	0.8380
YOLOv8	0.8507
Our model	0.8562

Attention mechanisms module	mAP_0.5
CA	0.8543
SE	0.8498
CBAM	0.8277

Algorithm	CA	Deformable convolution	Improved SWA	$ P $/%	$ R $/%	mAP_0.5
Origin				85.75	80.72	0.8183
imp1	+			87.40	83.05	0.8543
imp2		+		88.17	81.43	0.8290
imp3			+	86.47	81.95	0.8295
imp4	+	+	+	90.17	83.97	0.8562

Improved YOLOv5-based radar object detection

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 0

编辑推荐

Metrics

本文评价