改进YOLOv5的SAR图像舰船目标检测

doi:10.3778/j.issn.1002-8331.2108-0308

摘要/Abstract

摘要： 近年来针对合成孔径雷达（synthetic aperture radar，SAR）图像中缺乏颜色和纹理细节的舰船检测技术在深度学习领域中得到了广泛研究，利用深度学习技术可以有效避免传统的复杂特征设计，并且检测精度得到极大改善。针对舰船目标检测框具有高长宽比和密集排列问题，提出一种基于改进YOLOv5的目标检测方法。该方法针对舰船目标检测框特点将检测框长宽作为参数进行综合考虑并对损失函数进行曲线优化，并结合坐标注意力机制（coordinate attention，CA），在模型轻量化的同时实现对舰船目标检测的高速与高精度并存。实验结果表明：相比原YOLOv5方法，该方法的检测精度由原来的92.3%提升到96.7%，mAP（mean average precision）指标由原来的92.5%提升到97.2%，明显优于对比方法。通过改进检测框损失函数和特征提取方式，提高对SAR图像中舰船目标的检测效果。

关键词: 合成孔径雷达（SAR）, YOLOv5, 舰船检测, 坐标注意力机制

Abstract: In recent years, the ship detection technology for the lack of color and texture details in synthetic aperture radar（SAR） images has been extensively studied in the field of deep learning. The use of deep learning technology can effectively avoid traditional complex feature design, and the accuracy of detection is greatly improved. For the problems of high aspect ratio and dense arrangement of ship targets detection, a target detection method based on improved YOLOv5 is proposed. According to the characteristics of ship targets detection, the length and width of detection are taken into comprehensive consideration and the loss function curve is optimized, and the coordinate attention mechanism（CA） is combined to achieve high-speed and high-precision detection of ship targets while the model is lightweight. The experimental results show that：Compared with the original YOLOv5 method, the detection accuracy of this method is increased from 92.3% to 96.7%, the mAP index is increased from 92.5% to 97.2%, which is significantly better than the comparison method. By improving the detection frame loss function and feature extraction methods, the detection effect of ship targets in SAR images is improved.

Key words: synthetic aperture radar（SAR）, YOLOv5, ship detection, coordinate attention mechanism

谭显东, 彭辉. 改进YOLOv5的SAR图像舰船目标检测[J]. 计算机工程与应用, 2022, 58(4): 247-254.

TAN Xiandong, PENG Hui. Improved YOLOv5 Ship Target Detection in SAR Image[J]. Computer Engineering and Applications, 2022, 58(4): 247-254.

参考文献

[1] ZHANG T W，ZHANG X L.High-speed ship detection in SAR images based on a grid convolutional neural network[J].Remote Sensing，2019，11（10）：1206.
[2] WACKERMAN C C，FRIEDMAN K S，PICHEL W G，et al.Automatic detection of ships in RADARSAT-1 SAR imagery[J].Canadian Journal of Remote Sensing，2001，27（5）：568-577.
[3] VACHON P W，THOMAS S J，CRANTON J，et al.Validation of ship detection by the RADARSAT synthetic aperture-radar and the ocean monitoring workstation[J].Canadian Journal of Remote Sensing，2000，26（3）：200-212.
[4] KAPUR J N，SAHOO P K，WONG A K C.A new method of gray-level picture thresholding using the entropy of the histogram[J].Computer Vision Graphics and Image Processing，1985，29（2）：273-285.
[5] WARDLOW B D，EGBERT S L，KASTENS J H.Analysis of time-series MODIS 250 m vegetation index data for crop classification in the U.S. Central Great Plains[J].Remote Sensing of Environment，2007，108（3）：290-310.
[6] WANG S G，WANG M，YANG S Y，et al.New hierarchical saliency filtering for fast ship detection in high-resolution SAR images[J].IEEE Transactions on Geoscience and Remote Sensing，2017，55（1）：351-362.
[7] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[8] GIRSHICK R.Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision，2015：1440-1448.
[9] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（6）：1137-1149.
[10] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition，2015：6517-6525.
[11] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multi-box detector[C]//14th European Conference on Computer Vision.Cham：Springer，2016：21-37.
[12] LIN T Y，GOYAL P，GIRSHICK R，et al.Focal loss for dense object detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42（2）：318-327.
[13] 顾佼佼，李炳臻，刘克，等.基于改进Faster R-CNN的红外舰船目标检测算法[J].红外技术，2021，43（2）：170-178.
GU J J，LI B Z，LIU K，et al.Infrared ship target detection algorithm based on improved Faster R-CNN[J].Infrared Technology，2021，43（2）：170-178.
[14] 张晓玲，张天文，师君，等.基于深度分离卷积神经网络的高速高精度SAR舰船检测[J].雷达学报，2019，8（6）：841-851.
ZHANG X L，ZHANG T W，SHI J，et al.High speed and high precision SAR ship detection based on deep separation convolutional neural network[J].Journal of Radar Science，2019，8（6）：841-851.
[15] WOO S，PARK J，LEE J Y，et al.CBAM：convolutional block attention module[C]//15th European Conference on Computer Vision，Munich，2018：3-19.
[16] 张麒麟，林清平，肖蕾.改进YOLOv5的航拍图像识别算法[J].长江信息通信，2021，34（3）：73-76.
ZHANG Q L，LIN Q P，XIAO L.Improved YOLOv5 aerial image recognition algorithm[J].Yangtze River Information Communication，201，34（3）：73-76.
[17] ZHENG Z，WANG P，LIU W，et al.Distance-IoU loss：faster and better learning for bounding box regression[C]//34th AAAI Conference on Artificial Intelligence，2020：12993-13000.
[18] 马永康，刘华，凌成星，等.基于改进YOLOv5的红树林单木目标检测研究[J/OL].激光与光电子学进展（2021-08-16）[2021-08-25].http：//kns.cnki.net/kcms/detail/31.1690.TN.
20210816.1533.054.html.
MA Y K，LIU H，LING C X，et al.Mangrove ChanMu target detection based on improved YOLOv5 research[J/OL].Laser and Optoelectronics Progress（2021-08-16）[2021-08-25].http：//kns.cnki.net/kcms/detail/31.1690.TN.20210816.1533.
054.html.
[19] STERGIOU A，POPPE R，KALLIATAKIS G.Refining activation downsampling with Softpool[J].arXiv：2101.00440，2021.
[20] HOU Q，ZHOU D，FENG J.Coordinate attention for efficient mobile network design[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：13713-13722.
[21] BOCHKOVSKIY A，WANG C Y，LIAO H.YOLOv4：optimal speed and accuracy of object detection[J].arXiv：2004.
10934，2020.
[22] WANG C Y，LIAO H Y M，WU Y H，et al.CSPNet：a new backbone that can enhance learning capability of CNN[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops，2020：390-391.
[23] REZATOFIGHI H，TSOI N，GWAK J Y，et al.Generalized intersection over union：a metric and a loss for bounding box regression[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：658-666.
[24] ZHANG Y F，REN W，ZHANG Z，et al.Focal and efficient IOU loss for accurate bounding box regression[J].arXiv：2101.08158，2021.
[25] 李健伟，曲长文，彭书娟.基于级联CNN的SAR图像舰船目标检测算法[J].控制与决策，2019，34（10）：2191-2197.
LI J W，QU C W，PENG S J.A ship detection method based on cascade CNN in SAR images[J].Control and Decision，2019，34（10）：2191-2197.