Improved YOLOv4-Tiny Fast Target Detection Method for SAR Image

doi:10.3778/j.issn.1002-8331.2203-0507

Abstract

Abstract: Nowadays, the target detection technology for synthetic aperture radar（SAR） images has been widely researched and applied, and the deep learning method can be effectively applied to ship detection under complex background features, and the detection accuracy and speed are also has certain improvement. In order to further improve the detection speed and accuracy in the SAR ship target scene, a fast detection method for a single type of target in SAR ship images is proposed. The YOLOv4-Tiny network is modified by applying depthwise separable convolution, and on top of this, improved channel and spatial attention mechanisms are added to improve accuracy. Taking the input image size of 608×608 as the benchmark, compared with the YOLOv4-Tiny model, the experimental results show that the detection accuracy of the method in this paper can reach 93.91%, the detection speed can reach 42.8 frame/s, and the model is improved by 11.7 frame/s. The weight is only 2.17 MB in size, which is about 1/10 of that of YOLOv4-Tiny. The model is conducive to the deployment of hardware on the board, and meets the needs of fast detection scenarios for ship targets in SAR images.

Key words: synthetic aperture radar, YOLOv4-Tiny, ship detection, depth separable convolution, attention mechanism

摘要： 如今面向合成孔径雷达（synthetic aperture radar，SAR）图像的目标检测技术已得到广泛的研究与应用，并通过深度学习的方法可有效应用于复杂背景特征下的舰船检测，检测精度和速度也具有一定的效果。为提高在SAR舰船目标场景中的检测速度和精度，提出了一种可用于SAR舰船图像的单类目标快速检测方法，主要通过应用深度可分离卷积修改YOLOv4-Tiny网络，并在此基础上添加经改进的通道和空间注意力机制来提升精度。以输入图像尺寸为608×608为基准，相比于YOLOv4-Tiny模型，实验结果表明，该方法的检测精度可达93.91%，检测速度可达42.8 frame/s，提高了11.7 frame/s，模型权重仅有2.17 MB大小，约为YOLOv4-Tiny的1/10。模型利于硬件上板部署，满足SAR图像舰船目标的快速检测场景需求。

关键词: 合成孔径雷达, YOLOv4-Tiny, 舰船检测, 深度可分离卷积, 注意力机制

ZHANG Kuo, CHEN Zhangjin, ZHANG Yan. Improved YOLOv4-Tiny Fast Target Detection Method for SAR Image[J]. Computer Engineering and Applications, 2023, 59(14): 209-216.

张廓, 陈章进, 张岩. 改进YOLOv4-Tiny的SAR图像目标快速检测方法[J]. 计算机工程与应用, 2023, 59(14): 209-216.

References

[1] ZHANG T，ZHANG X，SHI J，et al.Balance scene learning mechanism for offshore and inshore ship detection in SAR images[J].IEEE Geoscience and Remote Sensing Letters，2020，19：1-5.
[2] SEMPREVIVA A M，BARTHELMIE R J，PRYOR S C.Review of methodologies for offshore wind resource assessment in European seas[J].Surveys in Geophysics，2008，29（6）：471-497.
[3] GUI Y，LI X，XUE L.A multilayer fusion light-head detector for SAR ship detection[J].Sensors，2019，19（5）：1124.
[4] 李晨瑄，胥辉旗，钱坤，等.基于深度学习的舰船目标检测技术综述[J].兵器装备工程学报，2021，42（12）：57-63.
LI C X，XU H Q，QIAN K，et al.Survey of ship detection technology based on deep learning[J].Journal of Ordnance and Equipment Engineering，2021，42（12）：57-63.
[5] 王迪聪，白晨帅，邬开俊.基于深度学习的视频目标检测综述[J].计算机科学与探索，2021，15（9）：1563-1577.
WANG D C，BAI C S，WU K J.Survey of video object detection based on deep learning[J].Journal of Frontiers of Computer Science and Technology，2021，15（9）：1563-1577.
[6] DALAL N，TRIGGS B.Histograms of oriented gradients for human detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition（CVPR’05），2005：886-893.
[7] OJALA T，PIETIKAINEN M，MAENPAA T.Multiresolution gray-scale and rotation invariant texture classification with local binary patterns[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2002，24（7）：971-987.
[8] LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision，2004，60（2）：91-110.
[9] VIOLA P，JONES M.Rapid object detection using a boosted cascade of simple features[C]//Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2001.
[10] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[11] 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 & Machine Intelligence，2017，39（6）：1137-1149.
[12] HE K，GKIOXARI G，DOLLáR P，et al.Mask R-CNN[J].IEEE Transactions on Pattern Analysis & Machine Intelligence，2017：2961-2969.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision，2016：21-37.
[14] ZHOU X，WANG D，KR?HENBüHL P.Objects as points[J].arXiv：1904.07850，2019.
[15] TAN M，LE Q.EfficientNet：rethinking model scaling for convolutional neural networks[C]//International Conference on Machine Learning，2019：6105-6114.
[16] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[17] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：7263-7271.
[18] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[19] BOCHKOVSKIY A，WANG C Y，LIAO H.YOLOv4：optimal speed and accuracy of object detection[J].arXiv：2004.
10934，2020.
[20] EVERINGHAM M，GOOL L V，WILLIAMS C，et al.The pascal visual object classes（VOC） challenge[J].International Journal of Computer Vision，2010，88（2）：303-338.
[21] LIN T Y，MAIRE M，BELONGIE S，et al.Microsoft coco：common objects in context[C]//European Conference on Computer Vision.Cham：Springer，2014：740-755.
[22] 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.
[23] 董颖博.基于深度学习的 SAR 图像船舶检测与分类方法研究[D].北京：中国科学院大学（中国科学院空天信息创新研究院），2020：22-28.
DONG Y B.Research on ship detection and classification method of SAR image based on deep learning[D].Beijing：University of Chinese Academy of Sciences（Institute of Aerospace Information Innovation，Chinese Academy of Sciences），2020：22-28.
[24] 张磊，洪星，王岳环，等.高分辨率遥感图像投影分析的靠岸舰船检测[J].中国图象图形学报，2018，23（9）：1424-1432.
ZHANG L，HONG X，WANG Y H，et al.Inshore ship detection in high-resolution remote sensing image using projection analysis[J].Journal of Image and Graphics，2018，23（9）：1424-1432.
[25] 张欣，张永强，何斌，等.基于YOLOv4-Tiny的遥感图像飞机目标检测技术研究[J].光学技术，2021，47（3）：344-351.
ZHANG X，ZHANG Y Q，HE B，et al.Research on remote sensing image aircraft target detection techonlogy based on YOLOv4-tiny[J].Optical Technique，2021，47（3）：344-351.
[26] 吴靖，韩禄欣，沈英，等.基于改进YOLOv4-Tiny的无人机航拍目标检测[J].电光与控制，2022，29（12）：112-117.
WU J，HAN L X，SHEN Y，et al.UAV aerial target detection based on improved YOLOv4-tiny[J].Electronics Optics & Control，2022，29（12）：112-117.
[27] SANDLER M，HOWARD A，ZHU M，et al.Mobilenetv2：inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：4510-4520.
[28] HE K，ZHANG X，REN S，et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE Transactions on Pattern Analysis & Machine Intelligence，2014，37（9）：1904-1916.
[29] WOO S，PARK J，LEE J Y，et al.CBAM：convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision（ECCV），2018：3-19.
[30] WANG Q，WU B，ZHU P，et al.ECA-Net：efficient channel attention for deep convolutional neural networks[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR）.Seattle，USA：IEEE，2020：11531-11539.
[31] ZHANG T，ZHANG X.High-speed ship detection in SAR images based on a grid convolutional neural network[J].Remote Sensing，2019，11（10）：1206.
[32] JIE H，LI S，GANG S，et al.Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：7132-7141.