1.四川轻化工大学 自动化与信息工程学院, 四川 自贡 643000
2.人工智能四川省重点实验室, 四川 宜宾 644000
[ "周晴(1996—),女,江苏徐州人,硕士研究生,2020年于南京师范大学中北学院获得学士学位,主要从事图像处理、目标检测、目标识别等方面的研究。E-mail:1071432755@qq.com" ]
[ "谭功全(1970—),男,四川自贡人,硕士,教授,1999年于电子科技大学获得硕士学位,主要从事控制理论与控制科学等方面的研究。E-mail:tgq77@126.com" ]
扫 描 看 全 文
周晴, 谭功全, 尹宋麟, 等. 改进YOLOv5s的道路目标检测算法[J]. 液晶与显示, 2023,38(5):680-690.
ZHOU Qing, TAN Gong-quan, YIN Song-lin, et al. Road object detection algorithm based on improved YOLOv5s[J]. Chinese Journal of Liquid Crystals and Displays, 2023,38(5):680-690.
周晴, 谭功全, 尹宋麟, 等. 改进YOLOv5s的道路目标检测算法[J]. 液晶与显示, 2023,38(5):680-690. DOI: 10.37188/CJLCD.2022-0257.
ZHOU Qing, TAN Gong-quan, YIN Song-lin, et al. Road object detection algorithm based on improved YOLOv5s[J]. Chinese Journal of Liquid Crystals and Displays, 2023,38(5):680-690. DOI: 10.37188/CJLCD.2022-0257.
针对目前主流的目标检测算法存在模型参数过大、不能很好地移植到移动设备端之中应用于辅助驾驶这一问题,本文提出了一种改进YOLOv5s的目标检测算法。首先,将YOLOv5s算法的主干网络CSPDarknet替换为轻量化网络模型MobileNet-V3,解决了网络模型较大、参数较多的问题,减少了网络的深度并提升了数据推断速度;其次,对特征提取网络采用加权双向特征金字塔结构Bi-FPN加强特征提取,融合多尺度特征进而扩大感受野;最后,对损失函数进行优化,使用CIoU为边界框回归损失函数,改善模型原始GIoU收敛速度较慢问题,使预测框更加符合于真实框,同时降低网络训练难度。实验结果表明,改进后算法在KITTI数据集上的mAP相较于YOLOv5s、SSD、YOLOv3、YOLOv4_tiny分别提升了4.4、15.7、12.4、19.6,模型大小相较于YOLOv5s与轻量级网络YOLOv4_tiny分别减少了32.4 MB和21 MB,同时检测速度分别提升了17.6%与43%。本文改进后的算法满足模型小、精确度高的要求,为辅助驾驶中道路目标检测提升检测速度与精度提供了一种解决方案。
Aiming at the problem that the model parameters of the current mainstream target detection algorithms are too large and cannot be transplanted to mobile devices and applied to assisted driving, this paper proposes an improved YOLOv5s target detection algorithm. Firstly, CSPDarknet, the backbone network of YOLOv5s algorithm, is replaced by MobileNet-V3, a lightweight network model, which solves the problem of large network model and many parameters, reduces the network depth and improves the data inference speed. Secondly, a weighted bidirectional feature pyramid structure Bi-FPN is used to enhance feature extraction, and multi-scale features are integrated to expand the receptive field. Finally, the loss function is optimized and CIoU is used as the boundary box regression loss function to improve the slow convergence speed of the original GIoU model, so that the prediction box is more consistent with the real box, and at the same time reduce the difficulty of network training. Experimental results show that compared with YOLOv5s, SSD, YOLOv3 and YOLOv4_tiny, the mAP of the improved algorithm on KITTI dataset is improved by 4.4, 15.7, 12.4 and 19.6, respectively. Compared with YOLOv5s and lightweight network YOLOv4_tiny, the model size is reduced by 32.4 MB and 21 MB respectively, and the detection speed is improved by 17.6% and 43% respectively. The improved algorithm meets the requirements of small model and high accuracy, and provides a solution for improving detection speed and accuracy of road target detection in assisted driving.
MobileNetV3目标检测YOLOv5特征提取CIoU
MobileNetV3object detectionYOLOv5feature extractionCIoU
ZHAO Z Q, ZHENG P, XU S T, et al. Object detection with deep learning: a review [J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(11): 3212-3232. doi: 10.1109/tnnls.2018.2876865http://dx.doi.org/10.1109/tnnls.2018.2876865
YANG G H, FENG W, JIN J T, et al. Face mask recognition system with YOLOv5 based on image recognition [C]//2020 IEEE 6th International Conference on Computer and Communications (ICCC). Chengdu: IEEE, 2020: 1398-1404. doi: 10.1109/iccc51575.2020.9345042http://dx.doi.org/10.1109/iccc51575.2020.9345042
LECUN Y, BENGIO Y, HINTON G. Deep learning [J]. Nature, 2015, 521(7553): 436-444. doi: 10.1038/nature14539http://dx.doi.org/10.1038/nature14539
郭晓静,隋昊达.改进YOLOv3在机场跑道异物目标检测中的应用[J].计算机工程与应用,2021,57(8):249-255.
GUO X J, SUI H D. Application of improved YOLOv3 in foreign object debris target detection on airfield pavement [J]. Computer Engineering and Applications, 2021, 57(8): 249-255. (in Chinese)
符惠桐,王鹏,李晓艳,等.面向移动目标识别的轻量化网络模型[J].西安交通大学学报,2021,55(7):124-131. doi: 10.7652/xjtuxb202107014http://dx.doi.org/10.7652/xjtuxb202107014
FU H T, WANG P, LI X Y, et al. Lightweight network model for moving object recognition [J]. Journal of Xi'an Jiaotong University, 2021, 55(7): 124-131. (in Chinese). doi: 10.7652/xjtuxb202107014http://dx.doi.org/10.7652/xjtuxb202107014
孔雅洁,张叶.引入高斯掩码自注意力模块的YOLOv3目标检测方法[J].液晶与显示,2022,37(4):539-548. doi: 10.37188/cjlcd.2021-0250http://dx.doi.org/10.37188/cjlcd.2021-0250
KONG Y J, ZHANG Y. YOLOv3 object detection method by introducing Gaussian mask self-attention module [J]. Chinese Journal of Liquid Crystals and Displays, 2022, 37(4): 539-548. (in Chinese). doi: 10.37188/cjlcd.2021-0250http://dx.doi.org/10.37188/cjlcd.2021-0250
宋建辉,王思宇,刘砚菊,等.基于改进FFRCNN网络的无人机地面小目标检测算法[J].电光与控制,2022,29(7):69-73,80. doi: 10.3969/j.issn.1671-637X.2022.07.013http://dx.doi.org/10.3969/j.issn.1671-637X.2022.07.013
SONG J H, WANG S Y, LIU Y J, et al. Ground small target detection algorithm of UAV based on improved FFRCNN network [J]. Electronics Optics & Control, 2022, 29(7): 69-73, 80. (in Chinese). doi: 10.3969/j.issn.1671-637X.2022.07.013http://dx.doi.org/10.3969/j.issn.1671-637X.2022.07.013
崔丽群,杨振忠,段天龙,等.复合先验的显著性目标检测方法[J/OL].激光与光电子学进展,1-15.http://kns.cnki.net/kcms/detail/31.1690.TN.20191106.1202.054.htmlhttp://kns.cnki.net/kcms/detail/31.1690.TN.20191106.1202.054.html.
CUI L Q, YANG Z Z, DUAN T L, et al. Significance target detection method based on composite prior [J/OL]. Progress in Laser and Optoelectronics, 1-15. http://kns.cnki.net/kcms/detail/31.1690.TN.20191106.1202.054.html.http://kns.cnki.net/kcms/detail/31.1690.TN.20191106.1202.054.html.(in Chinese)
KAVYASHREE P S P, EL-SHARKAWY M. Compressed MobileNetv3: a light weight variant for resource-constrained platforms [C]//2021 IEEE 11th Annual Computing and Communication Workshop and Conference (CCWC). Las Vegas: IEEE, 2021: 104-107. doi: 10.1109/ccwc51732.2021.9376113http://dx.doi.org/10.1109/ccwc51732.2021.9376113
LIN S D, ZHU K X, FENG C, et al. Align-Yolact: a one-stage semantic segmentation network for real-time object detection [J]. Journal of Ambient Intelligence and Humanized Computing, 2023, 14(2): 863-870. doi: 10.1007/s12652-021-03340-4http://dx.doi.org/10.1007/s12652-021-03340-4
张广世,葛广英,朱荣华,等.基于改进YOLOv3网络的齿轮缺陷检测[J].激光与光电子学进展,2020,57(12):121009. doi: 10.3788/lop57.121009http://dx.doi.org/10.3788/lop57.121009
ZHANG G S, GE G Y, ZHU R H, et al. Gear defect detection based on the improved YOLOv3 network [J]. Laser & Optoelectronics Progress, 2020, 57(12): 121009. (in Chinese). doi: 10.3788/lop57.121009http://dx.doi.org/10.3788/lop57.121009
ZHENG Z H, WANG P, LIU W, et al. Distance-IoU loss: faster and better learning for bounding box regression [C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence. New York: AAAI, 2020: 12993-13000. doi: 10.1609/aaai.v34i07.6999http://dx.doi.org/10.1609/aaai.v34i07.6999
杨政,邓赵红,罗晓清,等.利用ELM-AE和迁移表征学习构建的目标跟踪系统[J].计算机科学与探索,2022,16(7):1633-1648. doi: 10.3778/j.issn.1673-9418.2012028http://dx.doi.org/10.3778/j.issn.1673-9418.2012028
YANG Z, DENG Z H, LUO X Q, et al. Target tracking system constructed by ELM-AE and transfer representation learning [J]. Journal of Frontiers of Computer Science & Technology, 2022, 16(7): 1633-1648. (in Chinese). doi: 10.3778/j.issn.1673-9418.2012028http://dx.doi.org/10.3778/j.issn.1673-9418.2012028
曹志伟.车载图像行人检测关键技术研究[D].北京:北京邮电大学,2021.
CAO Z W. Research on key technologies of pedestrian detection in vehicle image [D]. Beijing: Beijing University of Posts and Telecommunications, 2021. (in Chinese)
葛俏,梁桥康,邹坤霖,等.基于轻量化网络与嵌入式系统的喷码检测[J].控制工程,2022,29(12):2349-2356.
GE Q, LIANG Q K, ZOU K L, et al. Detection of inkjet code quality based on lightweight network and embedded system [J]. Control Engineering of China, 2022, 29(12): 2349-2356. (in Chinese)
侯志强,韩崇昭.视觉跟踪技术综述[J].自动化学报,2006,32(4):603-617.
HOU Z Q, HAN C Z. A survey of visual tracking [J]. Acta Automatica Sinica, 2006,32(4): 603-617. (in Chinese)
王建林,付雪松,黄展超,等.改进YOLOv2卷积神经网络的多类型合作目标检测[J].光学 精密工程,2020,28(1):251-260. doi: 10.3788/ope.20202801.0251http://dx.doi.org/10.3788/ope.20202801.0251
WANG J L, FU X S, HUANG Z C, et al. Multi-type cooperative targets detection using improved YOLOv2 convolutional neural network [J]. Optics and Precision Engineering, 2020, 28(1): 251-260. (in Chinese). doi: 10.3788/ope.20202801.0251http://dx.doi.org/10.3788/ope.20202801.0251
UHRIG J, SCHNEIDER N, SCHNEIDER L, et al. Sparsity invariant CNNs [C]//2017 International Conference on 3D Vision (3DV). Qingdao: IEEE, 2017: 11-20. doi: 10.1109/3dv.2017.00012http://dx.doi.org/10.1109/3dv.2017.00012
ZUO C, QIAN J M, FENG S J, et al. Deep learning in optical metrology: a review [J]. Light: Science & Applications, 2022, 11(1): 39. doi: 10.1038/s41377-022-00714-xhttp://dx.doi.org/10.1038/s41377-022-00714-x
晏晓天,黄山.基于分组异构卷积的轻量级目标检测网络[J].计算机科学,2020,47(4):108-111. doi: 10.11896/jsjkx.190600067http://dx.doi.org/10.11896/jsjkx.190600067
YAN X T, HUANG S. Light-weight object detection network based on grouping heterogeneous convolution [J]. Computer Science, 2020, 47(4): 108-111. (in Chinese). doi: 10.11896/jsjkx.190600067http://dx.doi.org/10.11896/jsjkx.190600067
张俊蓉,徐长彬,唐明周,等.基于SSD改进的目标检测方法研究[J].激光与红外,2019,49(8):1019-1025. doi: 10.3969/j.issn.1001-5078.2019.08.020http://dx.doi.org/10.3969/j.issn.1001-5078.2019.08.020
ZHANG J R, XU C B, TANG M Z, et al. The improved target detection methods based on SSD network [J]. Laser & Infrared, 2019, 49(8): 1019-1025. (in Chinese). doi: 10.3969/j.issn.1001-5078.2019.08.020http://dx.doi.org/10.3969/j.issn.1001-5078.2019.08.020
0
浏览量
47
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构